[{"data":1,"prerenderedAt":2492},["ShallowReactive",2],{"blog-mvp-one-workflow-not-twelve-weeks":3,"related-topic-mvp-one-workflow-not-twelve-weeks":569,"related-recent-mvp-one-workflow-not-twelve-weeks":1210},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"metaTitle":10,"topic":11,"author":12,"authorProfile":13,"coverImg":14,"coverImgAlt":15,"toc":16,"published":16,"createdAt":17,"updatedAt":17,"keywords":18,"faq":25,"body":38,"_type":563,"_id":564,"_source":565,"_file":566,"_stem":567,"_extension":568},"/blog/mvp-one-workflow-not-twelve-weeks","blog",false,"","Your MVP Doesn't Need 12 Weeks. It Needs One Workflow.","Every agency quotes 12 weeks for an MVP, and they're not wrong. But you don't need a finished MVP to learn whether the idea works. Here's what one workflow in two weeks actually gets you.","MVP in 2 Weeks? What You Can Actually Ship That Fast","Product and Engineering","Tarun Bhukya","https://www.linkedin.com/in/tarun-kumar-bhukya-40168b85/","/images/blog/mvp-one-workflow.webp","One core workflow shipped in two weeks, ahead of a full MVP build",true,"2026-09-01",[19,20,21,22,23,24],"mvp in 2 weeks","how long does it take to build an mvp","can an agency deliver an mvp in 8 weeks","rapid mvp development","mvp development timeline","build mvp fast",[26,29,32,35],{"question":27,"answer":28},"Can you really build an MVP in 2 weeks?","You can get one core workflow into production in two weeks, deployed, with real authentication and real data, so actual users can use it. A complete MVP with every workflow takes longer, usually 6 to 12 weeks. Anyone promising your whole product in two weeks is describing a prototype, not something your users can rely on.",{"question":30,"answer":31},"How long does it take to build an MVP?","Six to twelve weeks is the honest range for a full MVP, and complex products with compliance or heavy integrations run longer. The more useful question is how soon you can put something in front of real users, and that can be about two weeks if you pick one workflow instead of trying to finish everything.",{"question":33,"answer":34},"Can an agency deliver an MVP in 4, 6 or 8 weeks?","It depends entirely on how many workflows the MVP contains, which is why a timeline quoted before anyone has seen your scope means very little. Ask what specifically is being cut to hit the date. If nobody can answer that, the date is a guess.",{"question":36,"answer":37},"What is rapid MVP development?","Shipping the smallest useful piece of your product quickly so real users can tell you whether the idea works, instead of building every feature before launch. Done properly it is one core workflow in production, then a decision about what to build next based on what people actually did with it.",{"type":39,"children":40,"toc":550},"root",[41,49,54,61,66,71,76,81,87,92,97,102,108,182,187,193,198,203,208,213,218,223,229,234,239,284,289,295,300,305,310,323,328,333,339,344,354,364,374,384,390,395,400,405,411,416,421,426,431,460,466,473,477,484,488,495,499,506,510,514],{"type":42,"tag":43,"props":44,"children":45},"element","p",{},[46],{"type":47,"value":48},"text","We build MVPs for a living. So when a founder asks how fast we can build one, the answer that makes us the most money is a big number with a lot of scope attached to it.",{"type":42,"tag":43,"props":50,"children":51},{},[52],{"type":47,"value":53},"Here's the smaller one instead.",{"type":42,"tag":55,"props":56,"children":58},"h2",{"id":57},"the-12-week-quote-isnt-a-lie",[59],{"type":47,"value":60},"The 12-week quote isn't a lie",{"type":42,"tag":43,"props":62,"children":63},{},[64],{"type":47,"value":65},"If three agencies told you 12 weeks, they were being straight with you. A full MVP, with every workflow your product needs, genuinely does take somewhere between 6 and 12 weeks. Complex products with payments and compliance take longer.",{"type":42,"tag":43,"props":67,"children":68},{},[69],{"type":47,"value":70},"Anyone promising your entire product in a few days is describing a prototype. It will demo beautifully and fall over the first week real people use it.",{"type":42,"tag":43,"props":72,"children":73},{},[74],{"type":47,"value":75},"So the timeline isn't the problem.",{"type":42,"tag":43,"props":77,"children":78},{},[79],{"type":47,"value":80},"The question is.",{"type":42,"tag":55,"props":82,"children":84},{"id":83},"youre-asking-how-fast-you-should-be-asking-how-soon-youll-know",[85],{"type":47,"value":86},"You're asking how fast. You should be asking how soon you'll know.",{"type":42,"tag":43,"props":88,"children":89},{},[90],{"type":47,"value":91},"\"How fast can you build my MVP\" and \"how soon will I know if this works\" sound like the same question. They aren't, and the gap between them is where most of the runway goes.",{"type":42,"tag":43,"props":93,"children":94},{},[95],{"type":47,"value":96},"Twelve weeks to a finished MVP means twelve weeks before a single real user touches anything. Twelve weeks of building on your best guess about what people want. If the guess was good, you lost nothing. If it wasn't, you spent your whole budget finding out.",{"type":42,"tag":43,"props":98,"children":99},{},[100],{"type":47,"value":101},"There's a version where you find out in two.",{"type":42,"tag":55,"props":103,"children":105},{"id":104},"the-short-answer",[106],{"type":47,"value":107},"The Short Answer",{"type":42,"tag":109,"props":110,"children":111},"table",{},[112,131],{"type":42,"tag":113,"props":114,"children":115},"thead",{},[116],{"type":42,"tag":117,"props":118,"children":119},"tr",{},[120,126],{"type":42,"tag":121,"props":122,"children":123},"th",{},[124],{"type":47,"value":125},"What you want",{"type":42,"tag":121,"props":127,"children":128},{},[129],{"type":47,"value":130},"Realistic timeline",{"type":42,"tag":132,"props":133,"children":134},"tbody",{},[135,153,169],{"type":42,"tag":117,"props":136,"children":137},{},[138,144],{"type":42,"tag":139,"props":140,"children":141},"td",{},[142],{"type":47,"value":143},"One core workflow, in production, real users can use it",{"type":42,"tag":139,"props":145,"children":146},{},[147],{"type":42,"tag":148,"props":149,"children":150},"strong",{},[151],{"type":47,"value":152},"About 2 weeks",{"type":42,"tag":117,"props":154,"children":155},{},[156,161],{"type":42,"tag":139,"props":157,"children":158},{},[159],{"type":47,"value":160},"A full MVP with every workflow",{"type":42,"tag":139,"props":162,"children":163},{},[164],{"type":42,"tag":148,"props":165,"children":166},{},[167],{"type":47,"value":168},"6 to 12 weeks",{"type":42,"tag":117,"props":170,"children":171},{},[172,177],{"type":42,"tag":139,"props":173,"children":174},{},[175],{"type":47,"value":176},"A clickable prototype that demos well",{"type":42,"tag":139,"props":178,"children":179},{},[180],{"type":47,"value":181},"Days, and it proves nothing",{"type":42,"tag":43,"props":183,"children":184},{},[185],{"type":47,"value":186},"Two weeks doesn't buy you a product. It buys you evidence.",{"type":42,"tag":55,"props":188,"children":190},{"id":189},"what-one-workflow-actually-means",[191],{"type":47,"value":192},"What \"one workflow\" actually means",{"type":42,"tag":43,"props":194,"children":195},{},[196],{"type":47,"value":197},"This is where people assume two weeks means something trivial. A login screen. A landing page with a form.",{"type":42,"tag":43,"props":199,"children":200},{},[201],{"type":47,"value":202},"It doesn't.",{"type":42,"tag":43,"props":204,"children":205},{},[206],{"type":47,"value":207},"For Eitoss we built a data analysis workflow for the industrial sector: over 40 parameters, comparison across time periods, tables and graphs, and comparison against other custom flows. That was two weeks, in production.",{"type":42,"tag":43,"props":209,"children":210},{},[211],{"type":47,"value":212},"Some more of our own numbers, so you can calibrate rather than take our word for it. In Formester, our own product, building a form meant dragging and dropping every element, and how long it took depended on the size of the form and the design work involved. We added AI so you describe what you want in plain language, including the logic and the layout, and the form gets built in about 15 seconds. That feature took five days. Payment gateway integration took under a week. Connecting to external apps like Google Sheets took under three days.",{"type":42,"tag":43,"props":214,"children":215},{},[216],{"type":47,"value":217},"None of that is a prototype. It's deployed, and people use it.",{"type":42,"tag":43,"props":219,"children":220},{},[221],{"type":47,"value":222},"So when we say one workflow, we mean a real one. The thing your product exists to do, working end to end, with proper authentication and real data behind it.",{"type":42,"tag":55,"props":224,"children":226},{"id":225},"what-doesnt-fit-in-two-weeks",[227],{"type":47,"value":228},"What doesn't fit in two weeks",{"type":42,"tag":43,"props":230,"children":231},{},[232],{"type":47,"value":233},"The honest boundary matters more than the promise, because the promise is worthless without it.",{"type":42,"tag":43,"props":235,"children":236},{},[237],{"type":47,"value":238},"Two weeks is not enough for your whole product. It's also not enough for:",{"type":42,"tag":240,"props":241,"children":242},"ul",{},[243,254,264,274],{"type":42,"tag":244,"props":245,"children":246},"li",{},[247,252],{"type":42,"tag":148,"props":248,"children":249},{},[250],{"type":47,"value":251},"Multi-role permission systems.",{"type":47,"value":253}," The moment you have admins, managers and users all seeing different things, the complexity is in the rules, not the screens.",{"type":42,"tag":244,"props":255,"children":256},{},[257,262],{"type":42,"tag":148,"props":258,"children":259},{},[260],{"type":47,"value":261},"Payments plus compliance.",{"type":47,"value":263}," Payments alone can be quick. Payments with the compliance work around them is a different job.",{"type":42,"tag":244,"props":265,"children":266},{},[267,272],{"type":42,"tag":148,"props":268,"children":269},{},[270],{"type":47,"value":271},"Native mobile alongside web.",{"type":47,"value":273}," Either one can fit. Both at once, in two weeks, can't.",{"type":42,"tag":244,"props":275,"children":276},{},[277,282],{"type":42,"tag":148,"props":278,"children":279},{},[280],{"type":47,"value":281},"Anything waiting on a slow third party.",{"type":47,"value":283}," If your workflow depends on an integration that takes three weeks to get approved, no amount of engineering speed fixes that.",{"type":42,"tag":43,"props":285,"children":286},{},[287],{"type":47,"value":288},"If what you need is on that list, we'll tell you before you commit, not halfway through. A two-week sprint that can't prove anything is worse than a twelve-week build that can.",{"type":42,"tag":55,"props":290,"children":292},{"id":291},"the-part-nobody-tells-you-shipping-isnt-the-milestone",[293],{"type":47,"value":294},"The part nobody tells you: shipping isn't the milestone",{"type":42,"tag":43,"props":296,"children":297},{},[298],{"type":47,"value":299},"Here's the thing that actually decides whether any of this was worth it.",{"type":42,"tag":43,"props":301,"children":302},{},[303],{"type":47,"value":304},"Plenty of startups build every feature on the roadmap, launch, and don't grow. Then they look at the usage data and find people only ever touched three of the twenty things that got built. All that money went into features nobody asked for. Nobody was lazy and nobody was stupid. They just built for twelve weeks without ever checking.",{"type":42,"tag":43,"props":306,"children":307},{},[308],{"type":47,"value":309},"The teams that get somewhere do the opposite. They ship one workflow, watch how people actually use it, and decide the next step from what they saw.",{"type":42,"tag":43,"props":311,"children":312},{},[313,315,321],{"type":47,"value":314},"So the milestone isn't \"the code shipped.\" It's ",{"type":42,"tag":316,"props":317,"children":318},"em",{},[319],{"type":47,"value":320},"people used it",{"type":47,"value":322},".",{"type":42,"tag":43,"props":324,"children":325},{},[326],{"type":47,"value":327},"That's the part we push hardest on, and it's the part most agencies skip, because a shipped feature is easy to invoice and a used feature is somebody else's problem. Once your workflow is live, put it in front of real users before you build anything else. What they do with it should decide your next set of features and your next timeline, not the list you wrote before anyone had touched the product.",{"type":42,"tag":43,"props":329,"children":330},{},[331],{"type":47,"value":332},"We've built and run our own products, so we can help you read what comes back. That's usually more useful than the building.",{"type":42,"tag":55,"props":334,"children":336},{"id":335},"what-to-ask-the-agency-quoting-you-12-weeks",[337],{"type":47,"value":338},"What to ask the agency quoting you 12 weeks",{"type":42,"tag":43,"props":340,"children":341},{},[342],{"type":47,"value":343},"Use these on us too.",{"type":42,"tag":43,"props":345,"children":346},{},[347,352],{"type":42,"tag":148,"props":348,"children":349},{},[350],{"type":47,"value":351},"\"What specifically are you cutting to hit that date?\"",{"type":47,"value":353}," A timeline quoted before anyone has seen your scope is a guess dressed up as a plan. If they can't name what's out, the date means nothing.",{"type":42,"tag":43,"props":355,"children":356},{},[357,362],{"type":42,"tag":148,"props":358,"children":359},{},[360],{"type":47,"value":361},"\"What can I put in front of real users first, and when?\"",{"type":47,"value":363}," If the answer is \"at the end,\" you're being asked to spend the whole budget before you learn anything.",{"type":42,"tag":43,"props":365,"children":366},{},[367,372],{"type":42,"tag":148,"props":368,"children":369},{},[370],{"type":47,"value":371},"\"What happens if the first thing we ship doesn't land?\"",{"type":47,"value":373}," You're listening for whether they have a way to change direction, or whether the plan only works if the original feature list was right.",{"type":42,"tag":43,"props":375,"children":376},{},[377,382],{"type":42,"tag":148,"props":378,"children":379},{},[380],{"type":47,"value":381},"\"Who owns the code, and from when?\"",{"type":47,"value":383}," The right answer is you, from day one. Not on final payment.",{"type":42,"tag":55,"props":385,"children":387},{"id":386},"when-two-weeks-is-the-wrong-thing-to-buy",[388],{"type":47,"value":389},"When two weeks is the wrong thing to buy",{"type":42,"tag":43,"props":391,"children":392},{},[393],{"type":47,"value":394},"If you already know people want this, skip it.",{"type":42,"tag":43,"props":396,"children":397},{},[398],{"type":47,"value":399},"Some founders come to us with real evidence already: a waiting list, a manual version of the service they've been running in spreadsheets, paying customers on a competitor's product. If you have that, a two-week sprint is buying information you already own. Go straight to the full build.",{"type":42,"tag":43,"props":401,"children":402},{},[403],{"type":47,"value":404},"Same if your product only makes sense once several workflows exist together. Some marketplaces need both sides live to prove anything. Most regulated products need the compliance layer before anyone can legally touch them. In those cases we'd rather scope a longer build properly than sell you two weeks that can't answer the question.",{"type":42,"tag":55,"props":406,"children":408},{"id":407},"how-we-work",[409],{"type":47,"value":410},"How we work",{"type":42,"tag":43,"props":412,"children":413},{},[414],{"type":47,"value":415},"We agree the scope before anyone writes code: which workflow, what \"done\" means, how many people it takes, what it costs. A complex workflow needs more people to land in two weeks, and that's our problem to solve, settled upfront rather than discovered halfway through.",{"type":42,"tag":43,"props":417,"children":418},{},[419],{"type":47,"value":420},"Our rate is $25 to $80 an hour depending on complexity and dependencies. A two-week sprint is roughly 80 hours at a team size of one, more when the workflow needs more people. That prices one workflow, not a whole product.",{"type":42,"tag":43,"props":422,"children":423},{},[424],{"type":47,"value":425},"You own everything from day one. Code, infrastructure, accounts.",{"type":42,"tag":43,"props":427,"children":428},{},[429],{"type":47,"value":430},"If you want, tell us the one workflow that matters most in your product and we'll tell you honestly whether it fits in two weeks, and what it would take if it doesn't. No commitment, and if the answer is that you should go build the full thing, we'll say that.",{"type":42,"tag":43,"props":432,"children":433},{},[434,443,445,451,453,459],{"type":42,"tag":435,"props":436,"children":440},"a",{"href":437,"rel":438},"https://calendar.app.google/gbT42VeCDd7ioXh79",[439],"nofollow",[441],{"type":47,"value":442},"Book a call",{"type":47,"value":444},", or read more about how we run ",{"type":42,"tag":435,"props":446,"children":448},{"href":447},"/services/mvp-development/",[449],{"type":47,"value":450},"MVP development",{"type":47,"value":452}," and ",{"type":42,"tag":435,"props":454,"children":456},{"href":455},"/mvp-development-for-startups/",[457],{"type":47,"value":458},"MVP development for startups",{"type":47,"value":322},{"type":42,"tag":55,"props":461,"children":463},{"id":462},"common-questions",[464],{"type":47,"value":465},"Common questions",{"type":42,"tag":43,"props":467,"children":468},{},[469],{"type":42,"tag":148,"props":470,"children":471},{},[472],{"type":47,"value":27},{"type":42,"tag":43,"props":474,"children":475},{},[476],{"type":47,"value":28},{"type":42,"tag":43,"props":478,"children":479},{},[480],{"type":42,"tag":148,"props":481,"children":482},{},[483],{"type":47,"value":30},{"type":42,"tag":43,"props":485,"children":486},{},[487],{"type":47,"value":31},{"type":42,"tag":43,"props":489,"children":490},{},[491],{"type":42,"tag":148,"props":492,"children":493},{},[494],{"type":47,"value":33},{"type":42,"tag":43,"props":496,"children":497},{},[498],{"type":47,"value":34},{"type":42,"tag":43,"props":500,"children":501},{},[502],{"type":42,"tag":148,"props":503,"children":504},{},[505],{"type":47,"value":36},{"type":42,"tag":43,"props":507,"children":508},{},[509],{"type":47,"value":37},{"type":42,"tag":511,"props":512,"children":513},"hr",{},[],{"type":42,"tag":43,"props":515,"children":516},{},[517],{"type":42,"tag":316,"props":518,"children":519},{},[520,522,528,530,536,537,543,544],{"type":47,"value":521},"Related reading: ",{"type":42,"tag":435,"props":523,"children":525},{"href":524},"/blog/mvp-development-cost/",[526],{"type":47,"value":527},"How Much Does MVP Development Cost in 2026?",{"type":47,"value":529}," | ",{"type":42,"tag":435,"props":531,"children":533},{"href":532},"/blog/when-not-to-build-an-mvp/",[534],{"type":47,"value":535},"When Not to Build an MVP",{"type":47,"value":529},{"type":42,"tag":435,"props":538,"children":540},{"href":539},"/blog/mvp-development-process/",[541],{"type":47,"value":542},"The MVP Development Process",{"type":47,"value":529},{"type":42,"tag":435,"props":545,"children":547},{"href":546},"/blog/scope-software-project/",[548],{"type":47,"value":549},"How to Scope a Software Project",{"title":7,"searchDepth":551,"depth":551,"links":552},2,[553,554,555,556,557,558,559,560,561,562],{"id":57,"depth":551,"text":60},{"id":83,"depth":551,"text":86},{"id":104,"depth":551,"text":107},{"id":189,"depth":551,"text":192},{"id":225,"depth":551,"text":228},{"id":291,"depth":551,"text":294},{"id":335,"depth":551,"text":338},{"id":386,"depth":551,"text":389},{"id":407,"depth":551,"text":410},{"id":462,"depth":551,"text":465},"markdown","content:blog:mvp-one-workflow-not-twelve-weeks.md","content","blog/mvp-one-workflow-not-twelve-weeks.md","blog/mvp-one-workflow-not-twelve-weeks","md",[570,870],{"_path":571,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":572,"description":573,"metaTitle":574,"topic":11,"author":12,"authorProfile":13,"coverImg":575,"coverImgAlt":576,"toc":16,"published":16,"createdAt":577,"updatedAt":577,"keywords":578,"faq":585,"body":598,"_type":563,"_id":867,"_source":565,"_file":868,"_stem":869,"_extension":568},"/blog/add-ai-to-your-product","How to Add AI to Your Product Without Wrecking It","Adding AI to a product you already have? A practical, honest guide to doing it without breaking what works, wasting money, or shipping a useless bolt-on chatbot.","How to Add AI to Your Existing Product (Without Wrecking It)","/images/blog/add-ai-to-your-product.webp","A practical guide to adding AI to an existing product without breaking what already works.","2026-08-14",[579,580,581,582,583,584],"how to add ai to my product","add ai to existing product","ai integration","adding ai features","ai integration services","add ai to existing app",[586,589,592,595],{"question":587,"answer":588},"How do I add AI to my existing product?","Start with a real user problem, not with the AI. Pick one place where AI clearly helps, ground it in your own data so it's accurate, add guardrails and testing before it touches users, ship that one feature, and measure it before you add more. The mistake is bolting on a generic chatbot nobody asked for.",{"question":590,"answer":591},"Will adding AI break my current app?","It can, if it's bolted on carelessly, which is exactly how a lot of AI features go wrong. Done properly, AI is added as a contained feature with its own guardrails and tests, so it doesn't destabilize what already works. The key is treating it like any other production feature, not a magic sprinkle.",{"question":593,"answer":594},"Do I need to rebuild my product to add AI?","Almost never. In most cases AI is added into your existing product without a rebuild: the feature, the data grounding, and the infrastructure to run it reliably. If someone tells you that adding AI means starting over, get a second opinion.",{"question":596,"answer":597},"How much does it cost to add an AI feature?","The build scales with scope, and there's an ongoing run cost too: roughly $0.10 to $0.90 per AI action, plus hosting and, if it uses your data, a RAG setup. A small, focused feature is inexpensive to run. See our breakdown of what an AI agent actually costs for the full picture.",{"type":39,"children":599,"toc":851},[600,605,610,620,626,631,636,642,647,659,665,689,694,700,705,717,723,728,741,747,752,757,763,768,781,787,793,797,802,806,811,815,820,824],{"type":42,"tag":43,"props":601,"children":602},{},[603],{"type":47,"value":604},"Adding AI to a product you already have is not the same as building something new. You have real users, real data, and real things that work today. The goal is to make it better, not to break the thing that's paying the bills. And that's exactly where a lot of AI features go wrong.",{"type":42,"tag":43,"props":606,"children":607},{},[608],{"type":47,"value":609},"So here's a practical, honest guide to adding AI to an existing product without wrecking it, from a team that does this for clients and inside our own SaaS.",{"type":42,"tag":43,"props":611,"children":612},{},[613,618],{"type":42,"tag":148,"props":614,"children":615},{},[616],{"type":47,"value":617},"The short version:",{"type":47,"value":619}," start with a real problem, not with the AI. Add it where it genuinely helps, ground it in your own data, put guardrails and testing around it before it touches users, ship one feature, and measure it before you add more. Small and useful beats big and broken.",{"type":42,"tag":55,"props":621,"children":623},{"id":622},"start-with-the-problem-not-the-ai",[624],{"type":47,"value":625},"Start With the Problem, Not the AI",{"type":42,"tag":43,"props":627,"children":628},{},[629],{"type":47,"value":630},"The most common mistake is adding AI because everyone else is, not because it solves anything. That's how you end up with a feature nobody uses and a bill every month.",{"type":42,"tag":43,"props":632,"children":633},{},[634],{"type":47,"value":635},"Before anything else, name the actual problem. Where do your users get stuck, waste time, or ask for help? AI is worth adding when it makes one of those moments clearly better. If you can't point at the problem, you're not ready to add the AI yet, and we'll tell you that honestly rather than build you something shiny and useless.",{"type":42,"tag":55,"props":637,"children":639},{"id":638},"avoid-the-bolt-on-trap",[640],{"type":47,"value":641},"Avoid the Bolt-On Trap",{"type":42,"tag":43,"props":643,"children":644},{},[645],{"type":47,"value":646},"The classic failed AI feature is a chatbot slapped in the corner of the screen, disconnected from everything, answering generic questions badly. It looks like \"we have AI now.\" It helps no one.",{"type":42,"tag":43,"props":648,"children":649},{},[650,652,657],{"type":47,"value":651},"The better approach is to build AI ",{"type":42,"tag":316,"props":653,"children":654},{},[655],{"type":47,"value":656},"into",{"type":47,"value":658}," the workflows people already use. Not a separate chatbot island, but the AI quietly doing the work at the exact moment it helps: summarizing the thing they're looking at, drafting the reply they were about to write, catching the error before they submit. Woven in, not bolted on.",{"type":42,"tag":55,"props":660,"children":662},{"id":661},"ground-it-in-your-own-data",[663],{"type":47,"value":664},"Ground It in Your Own Data",{"type":42,"tag":43,"props":666,"children":667},{},[668,670,675,677,681,683,687],{"type":47,"value":669},"Generic AI knows the internet. It doesn't know ",{"type":42,"tag":316,"props":671,"children":672},{},[673],{"type":47,"value":674},"your",{"type":47,"value":676}," product, ",{"type":42,"tag":316,"props":678,"children":679},{},[680],{"type":47,"value":674},{"type":47,"value":682}," customers, or ",{"type":42,"tag":316,"props":684,"children":685},{},[686],{"type":47,"value":674},{"type":47,"value":688}," rules, and when it guesses, it guesses wrong in ways that embarrass you.",{"type":42,"tag":43,"props":690,"children":691},{},[692],{"type":47,"value":693},"The fix is grounding the AI in your own data, usually with a setup called RAG, so its answers come from your actual knowledge instead of thin air. This is the difference between an AI feature that's genuinely useful and one that confidently tells your customer something untrue.",{"type":42,"tag":55,"props":695,"children":697},{"id":696},"put-guardrails-around-it-before-it-ships",[698],{"type":47,"value":699},"Put Guardrails Around It Before It Ships",{"type":42,"tag":43,"props":701,"children":702},{},[703],{"type":47,"value":704},"The moment an AI feature can touch customer data, send something, or take an action, it needs guardrails. What is it allowed to do? What happens when it's unsure? How do you stop it from doing something it shouldn't?",{"type":42,"tag":43,"props":706,"children":707},{},[708,710,716],{"type":47,"value":709},"You also need to actually test it, not just try it once and ship. An AI feature that works in your demo can fail in a hundred small ways with real users. Evaluating it before launch is what keeps a helpful feature from becoming a public mistake. This is the same discipline that AI-built apps usually skip, which is ",{"type":42,"tag":435,"props":711,"children":713},{"href":712},"/blog/where-ai-built-apps-break/",[714],{"type":47,"value":715},"where they break",{"type":47,"value":322},{"type":42,"tag":55,"props":718,"children":720},{"id":719},"design-for-the-running-cost",[721],{"type":47,"value":722},"Design for the Running Cost",{"type":42,"tag":43,"props":724,"children":725},{},[726],{"type":47,"value":727},"Unlike a normal feature, an AI feature keeps costing money every time someone uses it. Each AI action has a small cost, and it adds up with volume. That's not a reason to avoid AI. It's a reason to design for it: use the right-sized model for the job, don't call the AI when you don't need to, and know your numbers before you scale.",{"type":42,"tag":43,"props":729,"children":730},{},[731,733,739],{"type":47,"value":732},"If you want the full picture, we broke down ",{"type":42,"tag":435,"props":734,"children":736},{"href":735},"/blog/ai-agent-development-cost/",[737],{"type":47,"value":738},"what it actually costs to build and run an AI agent",{"type":47,"value":740},", and the same logic applies to any AI feature.",{"type":42,"tag":55,"props":742,"children":744},{"id":743},"ship-small-measure-then-expand",[745],{"type":47,"value":746},"Ship Small, Measure, Then Expand",{"type":42,"tag":43,"props":748,"children":749},{},[750],{"type":47,"value":751},"One AI feature done well beats ten half-working ones. Pick the single highest-value place, ship it properly, and watch what real users do with it. If it earns its keep, expand. If it doesn't, you've learned that cheaply instead of betting the roadmap on it.",{"type":42,"tag":43,"props":753,"children":754},{},[755],{"type":47,"value":756},"This is the honest way to add AI: not a big bang, but a real feature that works, that you can trust, and that you can build on.",{"type":42,"tag":55,"props":758,"children":760},{"id":759},"getting-help-with-it",[761],{"type":47,"value":762},"Getting Help With It",{"type":42,"tag":43,"props":764,"children":765},{},[766],{"type":47,"value":767},"Adding AI well takes two things at once: knowing your product and knowing how to build AI that survives production. That's exactly the pairing we bring. We've added AI into our own SaaS and delivered it for a client, so we know both the \"make it useful\" and the \"keep it from breaking\" side.",{"type":42,"tag":43,"props":769,"children":770},{},[771,773,779],{"type":47,"value":772},"If you're weighing where AI fits in your product, that's a good first conversation, and we'll be straight about whether it's worth it. See ",{"type":42,"tag":435,"props":774,"children":776},{"href":775},"/services/ai-development/",[777],{"type":47,"value":778},"AI development services",{"type":47,"value":780}," for how we work.",{"type":42,"tag":55,"props":782,"children":784},{"id":783},"frequently-asked-questions",[785],{"type":47,"value":786},"Frequently Asked Questions",{"type":42,"tag":788,"props":789,"children":791},"h3",{"id":790},"how-do-i-add-ai-to-my-existing-product",[792],{"type":47,"value":587},{"type":42,"tag":43,"props":794,"children":795},{},[796],{"type":47,"value":588},{"type":42,"tag":788,"props":798,"children":800},{"id":799},"will-adding-ai-break-my-current-app",[801],{"type":47,"value":590},{"type":42,"tag":43,"props":803,"children":804},{},[805],{"type":47,"value":591},{"type":42,"tag":788,"props":807,"children":809},{"id":808},"do-i-need-to-rebuild-my-product-to-add-ai",[810],{"type":47,"value":593},{"type":42,"tag":43,"props":812,"children":813},{},[814],{"type":47,"value":594},{"type":42,"tag":788,"props":816,"children":818},{"id":817},"how-much-does-it-cost-to-add-an-ai-feature",[819],{"type":47,"value":596},{"type":42,"tag":43,"props":821,"children":822},{},[823],{"type":47,"value":597},{"type":42,"tag":43,"props":825,"children":826},{},[827],{"type":42,"tag":316,"props":828,"children":829},{},[830,831,836,838,843,845,850],{"type":47,"value":521},{"type":42,"tag":435,"props":832,"children":833},{"href":775},[834],{"type":47,"value":835},"AI Development Services",{"type":47,"value":837},", ",{"type":42,"tag":435,"props":839,"children":840},{"href":735},[841],{"type":47,"value":842},"What It Actually Costs to Build a Production AI Agent",{"type":47,"value":844},", and ",{"type":42,"tag":435,"props":846,"children":847},{"href":712},[848],{"type":47,"value":849},"Where AI-Built Apps Actually Break",{"type":47,"value":322},{"title":7,"searchDepth":551,"depth":551,"links":852},[853,854,855,856,857,858,859,860],{"id":622,"depth":551,"text":625},{"id":638,"depth":551,"text":641},{"id":661,"depth":551,"text":664},{"id":696,"depth":551,"text":699},{"id":719,"depth":551,"text":722},{"id":743,"depth":551,"text":746},{"id":759,"depth":551,"text":762},{"id":783,"depth":551,"text":786,"children":861},[862,864,865,866],{"id":790,"depth":863,"text":587},3,{"id":799,"depth":863,"text":590},{"id":808,"depth":863,"text":593},{"id":817,"depth":863,"text":596},"content:blog:add-ai-to-your-product.md","blog/add-ai-to-your-product.md","blog/add-ai-to-your-product",{"_path":871,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":842,"description":872,"metaTitle":873,"topic":11,"author":12,"authorProfile":13,"coverImg":874,"coverImgAlt":875,"toc":16,"published":16,"createdAt":876,"updatedAt":876,"keywords":877,"faq":884,"body":897,"_type":563,"_id":1207,"_source":565,"_file":1208,"_stem":1209,"_extension":568},"/blog/ai-agent-development-cost","What does an AI agent really cost to build and run? An honest breakdown of build costs, per-action costs, hosting, and RAG, from a team that runs its own agents.","AI Agent Development Cost in 2026: What It Actually Costs","/images/blog/ai-agent-development-cost.webp","An honest breakdown of what it costs to build and run a production AI agent: build, per-action, hosting, and RAG.","2026-08-12",[878,879,880,881,882,883],"ai agent development cost","how much does an ai agent cost","cost to build an ai agent","ai agent pricing","cost of running an ai agent","ai agent hosting cost",[885,888,891,894],{"question":886,"answer":887},"How much does it cost to build an AI agent?","There's no single number, because the build is engineering time that scales with scope. A single-action agent is a small job. An agent that touches several systems, needs guardrails you can trust, and has to be reliable in production is a multi-week to multi-month build. On top of the build there are running costs: roughly $0.10 to $0.90 per action, cloud hosting from about $15 to $20 a month, and RAG from about $30 a month if it works from your data.",{"question":889,"answer":890},"What are the ongoing costs of running an AI agent?","Three main things. Model usage, charged per action, usually around $0.10 to $0.90 each and more for complex actions. Hosting the agent, which is just hosting a small private system, from roughly $15 to $20 a month in the cloud and up with scale. And if it uses your own data, a RAG setup starting around $30 a month that grows with your data volume.",{"question":892,"answer":893},"Can I run an AI agent cheaply?","Yes. You can self-host it. Our own finance agent, Freddy, runs on our own machine, private to our team, so its infrastructure cost is close to nothing. Costs rise as you move to the cloud, add more actions, or grow the data behind it, but a small internal agent can be very cheap to run.",{"question":895,"answer":896},"Do I even need a custom AI agent?","Not always. If a single prompt or a simple automation already does the job, a full agent is overkill and a waste of money. An agent earns its cost when it takes real, multi-step actions with consequences. We'll tell you honestly when a simpler tool would serve you better.",{"type":39,"children":898,"toc":1194},[899,904,913,919,924,929,953,958,964,969,1012,1017,1023,1028,1045,1062,1079,1084,1090,1095,1100,1105,1111,1116,1129,1133,1138,1142,1147,1151,1156,1160,1165,1169],{"type":42,"tag":43,"props":900,"children":901},{},[902],{"type":47,"value":903},"\"How much does an AI agent cost?\" is a fair question with an annoying answer: it depends. But \"it depends\" is only useful if someone tells you what it depends on. So here's the honest version, with real numbers where we can give them, from a team that builds agents and runs several of our own.",{"type":42,"tag":43,"props":905,"children":906},{},[907,911],{"type":42,"tag":148,"props":908,"children":909},{},[910],{"type":47,"value":617},{"type":47,"value":912}," the build is engineering time that scales with how much the agent does. On top of that, an agent has running costs most people forget: roughly $0.10 to $0.90 per action, cloud hosting from about $15 to $20 a month, and, if it uses your data, RAG from about $30 a month. You can also self-host it and run it for almost nothing, which is exactly what we do with our own finance agent.",{"type":42,"tag":55,"props":914,"children":916},{"id":915},"the-two-costs-people-forget-to-separate",[917],{"type":47,"value":918},"The Two Costs People Forget to Separate",{"type":42,"tag":43,"props":920,"children":921},{},[922],{"type":47,"value":923},"Most people ask about the build cost and stop there. But an AI agent is not like a website you build once and forget. It's a small system that keeps running, and running it has its own cost.",{"type":42,"tag":43,"props":925,"children":926},{},[927],{"type":47,"value":928},"So there are really two questions:",{"type":42,"tag":930,"props":931,"children":932},"ol",{},[933,943],{"type":42,"tag":244,"props":934,"children":935},{},[936,941],{"type":42,"tag":148,"props":937,"children":938},{},[939],{"type":47,"value":940},"What does it cost to build?",{"type":47,"value":942}," (a one-time engineering cost)",{"type":42,"tag":244,"props":944,"children":945},{},[946,951],{"type":42,"tag":148,"props":947,"children":948},{},[949],{"type":47,"value":950},"What does it cost to run, every month?",{"type":47,"value":952}," (an ongoing cost)",{"type":42,"tag":43,"props":954,"children":955},{},[956],{"type":47,"value":957},"Get clear on both before you budget, because a cheap build with expensive running costs, or the other way around, will surprise you later.",{"type":42,"tag":55,"props":959,"children":961},{"id":960},"what-drives-the-build-cost",[962],{"type":47,"value":963},"What Drives the Build Cost",{"type":42,"tag":43,"props":965,"children":966},{},[967],{"type":47,"value":968},"The build is engineering time, and it scales with scope. The things that move the number most:",{"type":42,"tag":240,"props":970,"children":971},{},[972,982,992,1002],{"type":42,"tag":244,"props":973,"children":974},{},[975,980],{"type":42,"tag":148,"props":976,"children":977},{},[978],{"type":47,"value":979},"How many actions it takes.",{"type":47,"value":981}," An agent that does one thing is a small job. An agent that does ten, in sequence, with decisions in between, is a much bigger one.",{"type":42,"tag":244,"props":983,"children":984},{},[985,990],{"type":42,"tag":148,"props":986,"children":987},{},[988],{"type":47,"value":989},"How many systems it touches.",{"type":47,"value":991}," Every tool or system the agent connects to (your email, your database, a payment provider, a CRM) is more to build and more to keep reliable.",{"type":42,"tag":244,"props":993,"children":994},{},[995,1000],{"type":42,"tag":148,"props":996,"children":997},{},[998],{"type":47,"value":999},"How much you need to trust it.",{"type":47,"value":1001}," An agent that touches money or customer data needs real guardrails, error handling, and testing. That reliability is where a lot of the serious engineering goes, and it's not optional once there are consequences.",{"type":42,"tag":244,"props":1003,"children":1004},{},[1005,1010],{"type":42,"tag":148,"props":1006,"children":1007},{},[1008],{"type":47,"value":1009},"Whether it needs your data.",{"type":47,"value":1011}," An agent that answers from your own knowledge needs a RAG setup, which adds to both the build and the running cost.",{"type":42,"tag":43,"props":1013,"children":1014},{},[1015],{"type":47,"value":1016},"This is why anyone who quotes you a flat price without asking what the agent actually does is guessing. A single-action agent can be a couple of weeks. A complex, multi-system agent you can trust in production is a multi-week to multi-month build.",{"type":42,"tag":55,"props":1018,"children":1020},{"id":1019},"the-running-costs-the-part-people-forget",[1021],{"type":47,"value":1022},"The Running Costs (The Part People Forget)",{"type":42,"tag":43,"props":1024,"children":1025},{},[1026],{"type":47,"value":1027},"Here's where we can give you real numbers, because we pay these ourselves.",{"type":42,"tag":43,"props":1029,"children":1030},{},[1031,1036,1038,1043],{"type":42,"tag":148,"props":1032,"children":1033},{},[1034],{"type":47,"value":1035},"Per action.",{"type":47,"value":1037}," Every action an agent takes has a small model cost, usually somewhere around ",{"type":42,"tag":148,"props":1039,"children":1040},{},[1041],{"type":47,"value":1042},"$0.10 to $0.90",{"type":47,"value":1044},", depending on which model you use and how complex the action is. Simple actions sit at the low end. A complex, multi-step action costs more. This adds up with volume, so an agent doing thousands of actions a day is a real line item, while one doing a few dozen is pocket change.",{"type":42,"tag":43,"props":1046,"children":1047},{},[1048,1053,1055,1060],{"type":42,"tag":148,"props":1049,"children":1050},{},[1051],{"type":47,"value":1052},"Hosting.",{"type":47,"value":1054}," The agent needs somewhere to run. Hosting the orchestrator in the cloud is really just hosting a small private system, which typically ",{"type":42,"tag":148,"props":1056,"children":1057},{},[1058],{"type":47,"value":1059},"starts around $15 to $20 a month",{"type":47,"value":1061}," and climbs with scale. You don't have to use the cloud, though. Our own finance agent, Freddy, runs on our own machine, private to our team, so its infrastructure cost is close to nothing. It creates invoices, sends emails, and runs the numbers, all self-hosted.",{"type":42,"tag":43,"props":1063,"children":1064},{},[1065,1070,1072,1077],{"type":42,"tag":148,"props":1066,"children":1067},{},[1068],{"type":47,"value":1069},"RAG (if it uses your data).",{"type":47,"value":1071}," If the agent works from your own knowledge base, that's a RAG setup, and its cost tracks your data volume. It can ",{"type":42,"tag":148,"props":1073,"children":1074},{},[1075],{"type":47,"value":1076},"start around $30 a month",{"type":47,"value":1078}," and grows as the data and usage grow.",{"type":42,"tag":43,"props":1080,"children":1081},{},[1082],{"type":47,"value":1083},"None of these are huge on their own. The point is to know they exist, so \"we built the agent\" doesn't quietly become a monthly bill you didn't plan for.",{"type":42,"tag":55,"props":1085,"children":1087},{"id":1086},"where-founders-overspend",[1088],{"type":47,"value":1089},"Where Founders Overspend",{"type":42,"tag":43,"props":1091,"children":1092},{},[1093],{"type":47,"value":1094},"The most expensive mistake isn't paying too much for an agent. It's building an agent you didn't need.",{"type":42,"tag":43,"props":1096,"children":1097},{},[1098],{"type":47,"value":1099},"If a single prompt, a simple automation, or an off-the-shelf tool already does the job, a full custom agent is overkill. An agent earns its cost when it takes real, multi-step actions with consequences, sends the email, updates the record, moves the money, and is trusted to do it correctly. If your task is simpler than that, build the simpler thing and keep the money.",{"type":42,"tag":43,"props":1101,"children":1102},{},[1103],{"type":47,"value":1104},"We'll say this plainly in a first conversation. We build agents for a living and we still tell people when they don't need one, because selling someone the expensive version of a problem they don't have is a bad way to start a relationship.",{"type":42,"tag":55,"props":1106,"children":1108},{"id":1107},"so-what-should-you-budget",[1109],{"type":47,"value":1110},"So What Should You Budget?",{"type":42,"tag":43,"props":1112,"children":1113},{},[1114],{"type":47,"value":1115},"Honestly, the best way to get a real number is a short conversation about what the agent actually needs to do. But as a rough shape: expect a one-time build that scales with scope, plus a modest monthly run cost made of per-action usage, hosting from around $15 to $20, and RAG from around $30 if you need it. A small internal agent can be genuinely cheap. A production agent that touches money and customers costs more to build well, and it's worth it, because the whole point is that you can trust it.",{"type":42,"tag":43,"props":1117,"children":1118},{},[1119,1121,1127],{"type":47,"value":1120},"If you want a real figure for your case, that's exactly the kind of thing we scope on a call. We've built and run our own agents, so we can tell you quickly whether an agent is even the right tool, and what it would actually cost. See ",{"type":42,"tag":435,"props":1122,"children":1124},{"href":1123},"/services/ai-agent-development/",[1125],{"type":47,"value":1126},"AI agent development",{"type":47,"value":1128}," for what we build.",{"type":42,"tag":55,"props":1130,"children":1131},{"id":783},[1132],{"type":47,"value":786},{"type":42,"tag":788,"props":1134,"children":1136},{"id":1135},"how-much-does-it-cost-to-build-an-ai-agent",[1137],{"type":47,"value":886},{"type":42,"tag":43,"props":1139,"children":1140},{},[1141],{"type":47,"value":887},{"type":42,"tag":788,"props":1143,"children":1145},{"id":1144},"what-are-the-ongoing-costs-of-running-an-ai-agent",[1146],{"type":47,"value":889},{"type":42,"tag":43,"props":1148,"children":1149},{},[1150],{"type":47,"value":890},{"type":42,"tag":788,"props":1152,"children":1154},{"id":1153},"can-i-run-an-ai-agent-cheaply",[1155],{"type":47,"value":892},{"type":42,"tag":43,"props":1157,"children":1158},{},[1159],{"type":47,"value":893},{"type":42,"tag":788,"props":1161,"children":1163},{"id":1162},"do-i-even-need-a-custom-ai-agent",[1164],{"type":47,"value":895},{"type":42,"tag":43,"props":1166,"children":1167},{},[1168],{"type":47,"value":896},{"type":42,"tag":43,"props":1170,"children":1171},{},[1172],{"type":42,"tag":316,"props":1173,"children":1174},{},[1175,1176,1181,1182,1188,1189,1193],{"type":47,"value":521},{"type":42,"tag":435,"props":1177,"children":1178},{"href":1123},[1179],{"type":47,"value":1180},"AI Agent Development",{"type":47,"value":837},{"type":42,"tag":435,"props":1183,"children":1185},{"href":1184},"/blog/build-with-ai-or-hire/",[1186],{"type":47,"value":1187},"Build It Yourself With AI or Hire a Developer?",{"type":47,"value":844},{"type":42,"tag":435,"props":1190,"children":1191},{"href":712},[1192],{"type":47,"value":849},{"type":47,"value":322},{"title":7,"searchDepth":551,"depth":551,"links":1195},[1196,1197,1198,1199,1200,1201],{"id":915,"depth":551,"text":918},{"id":960,"depth":551,"text":963},{"id":1019,"depth":551,"text":1022},{"id":1086,"depth":551,"text":1089},{"id":1107,"depth":551,"text":1110},{"id":783,"depth":551,"text":786,"children":1202},[1203,1204,1205,1206],{"id":1135,"depth":863,"text":886},{"id":1144,"depth":863,"text":889},{"id":1153,"depth":863,"text":892},{"id":1162,"depth":863,"text":895},"content:blog:ai-agent-development-cost.md","blog/ai-agent-development-cost.md","blog/ai-agent-development-cost",[1211,1577,1911,2275],{"_path":1212,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":1213,"description":1214,"topic":11,"author":12,"authorProfile":13,"coverImg":1215,"coverImgAlt":1216,"toc":16,"published":16,"createdAt":1217,"updatedAt":1217,"keywords":1218,"faq":1225,"body":1238,"_type":563,"_id":1574,"_source":565,"_file":1575,"_stem":1576,"_extension":568},"/blog/ai-for-software-development","AI for Software Development: How We Actually Use It (and Where It Breaks)","An honest, from-the-inside look at AI for software development from an AI-run dev agency: what it genuinely speeds up, where it breaks, and what that means for you.","/images/blog/ai-for-software-development.webp","How a real dev team uses AI for software development, what it speeds up and where it breaks.","2026-08-19",[1219,1220,1221,1222,1223,1224],"ai for software development","ai software development","ai tools for software development","using ai in software development","ai coding tools","will ai replace software developers",[1226,1229,1232,1235],{"question":1227,"answer":1228},"Will AI replace software developers?","Not for real production work, no. AI replaces a lot of the typing and the first-draft grunt work, which makes good engineers much faster. It does not replace the judgment that decides what to build, how to structure it so it lasts, and how to keep it secure and running. AI amplifies a strong engineer; it does not stand in for one.",{"question":1230,"answer":1231},"What is AI actually good at in software development?","Speed on the well-understood parts: boilerplate, first drafts of a feature, writing tests, routine refactors, explaining unfamiliar code, and research. Used well by someone who knows what good looks like, it compresses work that used to take months into weeks. We have taken builds from roughly four months to about one this way.",{"question":1233,"answer":1234},"Where does AI fail in software development?","The parts that decide whether software survives contact with real users: system architecture, the last 20% to production, security, and knowing what NOT to build. AI writes the feature in front of it, not the system around it, and it is confidently wrong often enough that unchecked AI code breaks in ways you find out about from your users.",{"question":1236,"answer":1237},"Should I use AI to build my software myself, or hire a team?","If you are still testing an idea or building something small, use AI yourself. If real users and money depend on it, you need people who can code and drive AI. We wrote a full honest guide to that decision separately (build it yourself with AI, or hire).",{"type":39,"children":1239,"toc":1561},[1240,1245,1250,1255,1264,1270,1275,1328,1333,1339,1344,1397,1402,1408,1413,1423,1433,1439,1444,1457,1462,1468,1473,1491,1495,1500,1505,1510,1514,1519,1524,1529,1539],{"type":42,"tag":43,"props":1241,"children":1242},{},[1243],{"type":47,"value":1244},"We are an AI-run software agency. We use AI in our own engineering every day, we build AI into products, and we run our own AI agents to operate the company. So when we talk about AI for software development, we have skin in the game, and you should read this with that in mind.",{"type":42,"tag":43,"props":1246,"children":1247},{},[1248],{"type":47,"value":1249},"Here is the honest version anyway, because the honest version is more useful than the hype or the doom.",{"type":42,"tag":43,"props":1251,"children":1252},{},[1253],{"type":47,"value":1254},"AI has genuinely changed how we build software. Not in the \"it does everything now\" way the headlines suggest, and not in the \"it is all a bubble\" way the skeptics claim. The truth is more specific and more useful: AI is excellent at some parts of software development and quietly terrible at others, and knowing the difference is the whole game.",{"type":42,"tag":43,"props":1256,"children":1257},{},[1258,1262],{"type":42,"tag":148,"props":1259,"children":1260},{},[1261],{"type":47,"value":617},{"type":47,"value":1263}," AI makes good engineers dramatically faster on the well-understood parts of the work, which is why a build that used to take us about four months now takes about one. It does not do the parts that decide whether software actually survives real users: architecture, the last stretch to production, security, and knowing what not to build. Those still need people.",{"type":42,"tag":55,"props":1265,"children":1267},{"id":1266},"what-ai-is-genuinely-good-at",[1268],{"type":47,"value":1269},"What AI Is Genuinely Good At",{"type":42,"tag":43,"props":1271,"children":1272},{},[1273],{"type":47,"value":1274},"Used by someone who knows what good looks like, AI is a real force multiplier. Where it earns its keep for us:",{"type":42,"tag":240,"props":1276,"children":1277},{},[1278,1288,1298,1308,1318],{"type":42,"tag":244,"props":1279,"children":1280},{},[1281,1286],{"type":42,"tag":148,"props":1282,"children":1283},{},[1284],{"type":47,"value":1285},"First drafts, fast.",{"type":47,"value":1287}," Give it a clear, well-scoped task and it produces a working first version in minutes instead of hours. You still review and shape it, but you start from something instead of a blank file.",{"type":42,"tag":244,"props":1289,"children":1290},{},[1291,1296],{"type":42,"tag":148,"props":1292,"children":1293},{},[1294],{"type":47,"value":1295},"The boring, necessary stuff.",{"type":47,"value":1297}," Boilerplate, wiring, config, repetitive refactors across a codebase. The work that is not hard, just slow. AI eats it.",{"type":42,"tag":244,"props":1299,"children":1300},{},[1301,1306],{"type":42,"tag":148,"props":1302,"children":1303},{},[1304],{"type":47,"value":1305},"Tests.",{"type":47,"value":1307}," It is genuinely good at drafting tests, which is the exact thing tired engineers skip. More tests, written faster, is a real quality win.",{"type":42,"tag":244,"props":1309,"children":1310},{},[1311,1316],{"type":42,"tag":148,"props":1312,"children":1313},{},[1314],{"type":47,"value":1315},"Understanding unfamiliar code.",{"type":47,"value":1317}," Point it at a codebase nobody remembers writing and it explains what is going on. That alone saves hours when you inherit or rescue a project.",{"type":42,"tag":244,"props":1319,"children":1320},{},[1321,1326],{"type":42,"tag":148,"props":1322,"children":1323},{},[1324],{"type":47,"value":1325},"Research and options.",{"type":47,"value":1327}," \"What are three ways to do X and the trade-offs\" is something it does well as a starting point for a real decision.",{"type":42,"tag":43,"props":1329,"children":1330},{},[1331],{"type":47,"value":1332},"Add these up and you get the four-months-to-one-month effect. That is not marketing. It is what happens when experienced engineers stop typing the easy 70% by hand and spend their time on the hard 30% instead.",{"type":42,"tag":55,"props":1334,"children":1336},{"id":1335},"where-ai-breaks",[1337],{"type":47,"value":1338},"Where AI Breaks",{"type":42,"tag":43,"props":1340,"children":1341},{},[1342],{"type":47,"value":1343},"This is the part the demos never show, and it is the part that actually matters.",{"type":42,"tag":240,"props":1345,"children":1346},{},[1347,1357,1367,1377,1387],{"type":42,"tag":244,"props":1348,"children":1349},{},[1350,1355],{"type":42,"tag":148,"props":1351,"children":1352},{},[1353],{"type":47,"value":1354},"Architecture and system design.",{"type":47,"value":1356}," AI writes the feature in front of it, not the system around it. It does not hold your whole product in its head, so left alone it produces code that works today and becomes a tangle you cannot build on in six months.",{"type":42,"tag":244,"props":1358,"children":1359},{},[1360,1365],{"type":42,"tag":148,"props":1361,"children":1362},{},[1363],{"type":47,"value":1364},"The last 20% to production.",{"type":47,"value":1366}," Getting something to demo is easy now. Getting it to stay up under real traffic, handle the failure cases, and not fall over when a hundred people use it at once is exactly as hard as it always was. AI is weakest right where it counts most.",{"type":42,"tag":244,"props":1368,"children":1369},{},[1370,1375],{"type":42,"tag":148,"props":1371,"children":1372},{},[1373],{"type":47,"value":1374},"Security.",{"type":47,"value":1376}," AI happily ships code with real, exploitable gaps, most often missing access controls that leave data exposed. It looks fine. It is not fine. You find out the expensive way if nobody who knows security reviewed it.",{"type":42,"tag":244,"props":1378,"children":1379},{},[1380,1385],{"type":42,"tag":148,"props":1381,"children":1382},{},[1383],{"type":47,"value":1384},"Knowing what not to build.",{"type":47,"value":1386}," AI will enthusiastically build whatever you ask for, including the wrong thing. It has no opinion about whether a feature is worth building. That judgment is still entirely on you (or the team you hire).",{"type":42,"tag":244,"props":1388,"children":1389},{},[1390,1395],{"type":42,"tag":148,"props":1391,"children":1392},{},[1393],{"type":47,"value":1394},"Confident wrongness.",{"type":47,"value":1396}," The failure mode that catches people out: AI is often wrong in the exact same confident tone it uses when it is right. If you cannot tell the difference, you cannot safely use it for anything that matters.",{"type":42,"tag":43,"props":1398,"children":1399},{},[1400],{"type":47,"value":1401},"The pattern across all of these is the same. AI is brilliant at the parts of software development that are well understood and bounded, and unreliable at the parts that need judgment, context, and responsibility.",{"type":42,"tag":55,"props":1403,"children":1405},{"id":1404},"so-what-does-this-actually-mean",[1406],{"type":47,"value":1407},"So What Does This Actually Mean?",{"type":42,"tag":43,"props":1409,"children":1410},{},[1411],{"type":47,"value":1412},"Two things, and they point in the same direction.",{"type":42,"tag":43,"props":1414,"children":1415},{},[1416,1421],{"type":42,"tag":148,"props":1417,"children":1418},{},[1419],{"type":47,"value":1420},"AI amplifies good engineers. It does not replace them.",{"type":47,"value":1422}," A strong engineer with AI is dramatically faster than they were a year ago. A non-engineer with AI can get surprisingly far, then hits a wall exactly where the judgment starts to matter. The value is not in the tool. It is in the person driving it who knows when the tool is wrong.",{"type":42,"tag":43,"props":1424,"children":1425},{},[1426,1431],{"type":42,"tag":148,"props":1427,"children":1428},{},[1429],{"type":47,"value":1430},"The advantage goes to whoever can do both.",{"type":47,"value":1432}," The people winning right now are the ones who know how to code and know how to drive AI. They get the speed of AI without inheriting the mess it leaves when nobody is checking. That combination is the actual skill of this era, and it is what you are really buying when you hire a team that has genuinely adopted AI rather than just talked about it.",{"type":42,"tag":55,"props":1434,"children":1436},{"id":1435},"how-we-use-ai-the-proof",[1437],{"type":47,"value":1438},"How We Use AI (the Proof)",{"type":42,"tag":43,"props":1440,"children":1441},{},[1442],{"type":47,"value":1443},"We do not just use AI on client work. We run our own company on it, which is the only reason we can speak to any of this honestly.",{"type":42,"tag":43,"props":1445,"children":1446},{},[1447,1449,1455],{"type":47,"value":1448},"We build AI into our own SaaS, ",{"type":42,"tag":435,"props":1450,"children":1452},{"href":1451},"/case-studies/formester/",[1453],{"type":47,"value":1454},"Formester",{"type":47,"value":1456},": an AI form builder, agentic workflows that take real actions after a submission, and a retrieval system grounded in real data. We run our own internal agents, including a finance agent that creates invoices and runs the numbers, a bug-fixing agent, and a code-review agent. And we deliver AI for clients, not just for ourselves.",{"type":42,"tag":43,"props":1458,"children":1459},{},[1460],{"type":47,"value":1461},"That day-to-day, in-production experience is what taught us where the line is between what AI can do and what it cannot. We learned it by shipping, not by reading about it.",{"type":42,"tag":55,"props":1463,"children":1465},{"id":1464},"the-honest-takeaway",[1466],{"type":47,"value":1467},"The Honest Takeaway",{"type":42,"tag":43,"props":1469,"children":1470},{},[1471],{"type":47,"value":1472},"AI for software development is neither magic nor a threat. It is a genuinely powerful tool that made building faster and left the hard parts exactly as hard. If you use it to skip the judgment, it will hurt you. If you use it to move faster on the work that does not need judgment, it is transformative.",{"type":42,"tag":43,"props":1474,"children":1475},{},[1476,1478,1483,1485,1490],{"type":47,"value":1477},"If you want AI built into your product properly, or you want a team that uses AI the way we do (for speed, with the judgment to catch it when it is wrong), that is what our ",{"type":42,"tag":435,"props":1479,"children":1480},{"href":775},[1481],{"type":47,"value":1482},"AI development",{"type":47,"value":1484}," work is. And if you are still deciding whether to build it yourself with AI or bring in help, we wrote an honest guide to exactly that: ",{"type":42,"tag":435,"props":1486,"children":1487},{"href":1184},[1488],{"type":47,"value":1489},"build it yourself with AI, or hire a developer",{"type":47,"value":322},{"type":42,"tag":55,"props":1492,"children":1493},{"id":783},[1494],{"type":47,"value":786},{"type":42,"tag":788,"props":1496,"children":1498},{"id":1497},"will-ai-replace-software-developers",[1499],{"type":47,"value":1227},{"type":42,"tag":43,"props":1501,"children":1502},{},[1503],{"type":47,"value":1504},"Not for real production work. AI replaces a lot of the typing and the first-draft grunt work, which makes good engineers much faster. It does not replace the judgment that decides what to build, how to structure it so it lasts, and how to keep it secure and running. AI amplifies a strong engineer; it does not stand in for one.",{"type":42,"tag":788,"props":1506,"children":1508},{"id":1507},"what-is-ai-actually-good-at-in-software-development",[1509],{"type":47,"value":1230},{"type":42,"tag":43,"props":1511,"children":1512},{},[1513],{"type":47,"value":1231},{"type":42,"tag":788,"props":1515,"children":1517},{"id":1516},"where-does-ai-fail-in-software-development",[1518],{"type":47,"value":1233},{"type":42,"tag":43,"props":1520,"children":1521},{},[1522],{"type":47,"value":1523},"The parts that decide whether software survives real users: system architecture, the last 20% to production, security, and knowing what not to build. AI writes the feature in front of it, not the system around it, and it is confidently wrong often enough that unchecked AI code breaks in ways you find out about from your users.",{"type":42,"tag":788,"props":1525,"children":1527},{"id":1526},"should-i-use-ai-to-build-my-software-myself-or-hire-a-team",[1528],{"type":47,"value":1236},{"type":42,"tag":43,"props":1530,"children":1531},{},[1532,1534,1538],{"type":47,"value":1533},"If you are still testing an idea or building something small, use AI yourself. If real users and money depend on it, you need people who can code and drive AI. We wrote a full honest guide to that decision: ",{"type":42,"tag":435,"props":1535,"children":1536},{"href":1184},[1537],{"type":47,"value":1489},{"type":47,"value":322},{"type":42,"tag":43,"props":1540,"children":1541},{},[1542],{"type":42,"tag":316,"props":1543,"children":1544},{},[1545,1546,1550,1551,1555,1556,1560],{"type":47,"value":521},{"type":42,"tag":435,"props":1547,"children":1548},{"href":775},[1549],{"type":47,"value":835},{"type":47,"value":837},{"type":42,"tag":435,"props":1552,"children":1553},{"href":712},[1554],{"type":47,"value":849},{"type":47,"value":844},{"type":42,"tag":435,"props":1557,"children":1558},{"href":1184},[1559],{"type":47,"value":1187},{"type":47,"value":322},{"title":7,"searchDepth":551,"depth":551,"links":1562},[1563,1564,1565,1566,1567,1568],{"id":1266,"depth":551,"text":1269},{"id":1335,"depth":551,"text":1338},{"id":1404,"depth":551,"text":1407},{"id":1435,"depth":551,"text":1438},{"id":1464,"depth":551,"text":1467},{"id":783,"depth":551,"text":786,"children":1569},[1570,1571,1572,1573],{"id":1497,"depth":863,"text":1227},{"id":1507,"depth":863,"text":1230},{"id":1516,"depth":863,"text":1233},{"id":1526,"depth":863,"text":1236},"content:blog:ai-for-software-development.md","blog/ai-for-software-development.md","blog/ai-for-software-development",{"_path":1578,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":1579,"description":1580,"metaTitle":1581,"topic":11,"author":12,"authorProfile":13,"coverImg":1582,"coverImgAlt":1583,"toc":16,"published":16,"createdAt":1584,"updatedAt":1584,"keywords":1585,"faq":1592,"body":1605,"_type":563,"_id":1908,"_source":565,"_file":1909,"_stem":1910,"_extension":568},"/blog/lovable-bubble-app-to-production","From a Lovable or Bubble App to a Real Product: The 5 Gaps to Close","Built on Lovable, Bubble, or another AI/no-code tool and ready for a real product? Here are the 5 gaps you have to close to migrate to production, without starting over.","From a Lovable or Bubble App to a Real Product (The 5 Gaps to Close)","/images/blog/lovable-bubble-app-to-production.webp","The five gaps to close when moving a Lovable or Bubble app to a real, production-grade product.","2026-08-16",[1586,1587,1588,1589,1590,1591],"lovable app to production","migrate from bubble to code","bubble app doesn't scale","move no-code app to custom","lovable to real product","take my no-code app to production",[1593,1596,1599,1602],{"question":1594,"answer":1595},"Can I turn my Lovable or Bubble app into a real product?","Yes. The tool got you a validated, working app, which is the hard part most people never reach. Turning it into a real product means closing a few specific gaps (ownership, scale, structure, security, and custom logic) usually by rebuilding it on code you own. You keep everything you learned; the workflows you built become the spec.",{"question":1597,"answer":1598},"Do I have to rebuild from scratch to move off Bubble or Lovable?","You rebuild the app on real code, but you don't start from scratch in the way that sounds. Every screen and workflow you figured out is a finished blueprint, so the thinking is done. It's a graduation, not a do-over, and it's usually faster than the first build because the product is already validated.",{"question":1600,"answer":1601},"Why won't my Bubble or Lovable app scale?","These tools are built for speed of building, not raw performance. Under real load, pages that were snappy with a little data start to lag, and you can't tune what's underneath because you don't control it. That performance ceiling is one of the most common reasons founders migrate to custom code.",{"question":1603,"answer":1604},"When should I migrate off a no-code or AI tool?","When you've hit the ceiling: workarounds stacked on workarounds, users feeling the performance, an integration the platform can't do, costs outpacing revenue, or realizing you can't take your app with you. If you're still validating, stay on the tool. Migration is for once it's working.",{"type":39,"children":1606,"toc":1892},[1607,1612,1617,1626,1632,1637,1650,1656,1661,1673,1679,1684,1689,1695,1700,1712,1718,1723,1728,1734,1739,1744,1750,1755,1788,1807,1811,1824,1829,1833,1838,1842,1847,1851,1856,1860,1865,1869],{"type":42,"tag":43,"props":1608,"children":1609},{},[1610],{"type":47,"value":1611},"You built it on Lovable, or Bubble, or one of the AI tools that turn a description into an app. It worked. You got users. And now you're looking at turning that into a real, production-grade product, and wondering if that means throwing everything away and starting over.",{"type":42,"tag":43,"props":1613,"children":1614},{},[1615],{"type":47,"value":1616},"It doesn't. The tool did the hardest part: it proved the thing works and people want it. What's left is closing a specific set of gaps between \"a working app on a platform\" and \"a real product you own and can grow.\" Here are the five, and how the move actually goes.",{"type":42,"tag":43,"props":1618,"children":1619},{},[1620,1624],{"type":42,"tag":148,"props":1621,"children":1622},{},[1623],{"type":47,"value":617},{"type":47,"value":1625}," you don't start over. You rebuild on code you own, and the workflows you already figured out become the blueprint. The gaps to close are ownership, scale, structure, security, and the custom logic the platform can't do. Migrate the load-bearing pieces first, keep the app live while you do it, and it's a graduation, not a do-over.",{"type":42,"tag":55,"props":1627,"children":1629},{"id":1628},"first-an-honest-check-should-you-migrate-yet",[1630],{"type":47,"value":1631},"First, an Honest Check: Should You Migrate Yet?",{"type":42,"tag":43,"props":1633,"children":1634},{},[1635],{"type":47,"value":1636},"Before the gaps, the honest question. If you're still figuring out whether people want this, don't migrate. Stay on the tool. It's cheaper and faster for learning, and migrating a product you haven't validated is effort you might throw away.",{"type":42,"tag":43,"props":1638,"children":1639},{},[1640,1642,1648],{"type":47,"value":1641},"Migration is for once it's working. If people are using it, paying for it, and asking for more, and you're starting to feel the walls, that's the signal. We wrote a full guide on ",{"type":42,"tag":435,"props":1643,"children":1645},{"href":1644},"/blog/no-code-vs-custom-development/",[1646],{"type":47,"value":1647},"when a no-code or AI-built app hits the ceiling",{"type":47,"value":1649}," if you want to pressure-test the timing first. Assuming you're there, here's what has to change.",{"type":42,"tag":55,"props":1651,"children":1653},{"id":1652},"gap-1-you-dont-own-the-code",[1654],{"type":47,"value":1655},"Gap 1: You Don't Own the Code",{"type":42,"tag":43,"props":1657,"children":1658},{},[1659],{"type":47,"value":1660},"On most of these platforms, you don't own source code. You own a configuration inside someone else's system. If the pricing changes, the platform pivots, or you simply outgrow it, your app doesn't come with you.",{"type":42,"tag":43,"props":1662,"children":1663},{},[1664,1666,1672],{"type":47,"value":1665},"Closing this gap is the whole point of migrating: getting your product onto code that's genuinely yours, that any team can pick up and build on. It's also why we care about this so much, having built our own product and written before about ",{"type":42,"tag":435,"props":1667,"children":1669},{"href":1668},"/blog/avoiding-vendor-lock-in/",[1670],{"type":47,"value":1671},"avoiding vendor lock-in",{"type":47,"value":322},{"type":42,"tag":55,"props":1674,"children":1676},{"id":1675},"gap-2-it-wont-scale-past-the-platforms-ceiling",[1677],{"type":47,"value":1678},"Gap 2: It Won't Scale Past the Platform's Ceiling",{"type":42,"tag":43,"props":1680,"children":1681},{},[1682],{"type":47,"value":1683},"These tools are optimized for building fast, not for raw performance. That's the right trade early on. But under real load, the app that felt instant with a little data starts to lag, and you can't fix it, because you don't control the engine underneath.",{"type":42,"tag":43,"props":1685,"children":1686},{},[1687],{"type":47,"value":1688},"A real product is built to stay fast as the data and traffic grow. That's a foundational choice you can only make in code you control.",{"type":42,"tag":55,"props":1690,"children":1692},{"id":1691},"gap-3-no-structure-or-tests-so-its-fragile",[1693],{"type":47,"value":1694},"Gap 3: No Structure or Tests, So It's Fragile",{"type":42,"tag":43,"props":1696,"children":1697},{},[1698],{"type":47,"value":1699},"Platform-built apps (and AI-built ones especially) tend to have no automated tests and no real structure underneath. That's why they break every time you add something. It works until it quietly doesn't, and you find out from a user.",{"type":42,"tag":43,"props":1701,"children":1702},{},[1703,1705,1710],{"type":47,"value":1704},"Rebuilding on real code means adding the safety net: structure that makes the app easy to change, and tests that catch a problem before your users do. This is the difference between an app you're scared to touch and one you can build on with confidence. (It's the same gap that makes ",{"type":42,"tag":435,"props":1706,"children":1707},{"href":712},[1708],{"type":47,"value":1709},"AI-built apps break in production",{"type":47,"value":1711},".)",{"type":42,"tag":55,"props":1713,"children":1715},{"id":1714},"gap-4-security-and-data-ownership",[1716],{"type":47,"value":1717},"Gap 4: Security and Data Ownership",{"type":42,"tag":43,"props":1719,"children":1720},{},[1721],{"type":47,"value":1722},"The gap you can't see. Platform-built apps often ship with weak access controls, the kind that let the wrong person reach data they shouldn't. And your customers' data lives inside a system you don't control.",{"type":42,"tag":43,"props":1724,"children":1725},{},[1726],{"type":47,"value":1727},"A production product handles security properly and keeps your data yours. Once real users trust you with real information, this stops being optional.",{"type":42,"tag":55,"props":1729,"children":1731},{"id":1730},"gap-5-the-custom-logic-the-platform-cant-do",[1732],{"type":47,"value":1733},"Gap 5: The Custom Logic the Platform Can't Do",{"type":42,"tag":43,"props":1735,"children":1736},{},[1737],{"type":47,"value":1738},"Eventually you need something the platform simply won't do: a specific integration, a custom workflow, a piece of the product that's genuinely yours and different. On a no-code tool you hit a hard wall. In custom code, that wall isn't there.",{"type":42,"tag":43,"props":1740,"children":1741},{},[1742],{"type":47,"value":1743},"Often this is the actual trigger for migrating: the next important thing you want to build is the one thing the tool can't.",{"type":42,"tag":55,"props":1745,"children":1747},{"id":1746},"how-the-migration-actually-goes",[1748],{"type":47,"value":1749},"How the Migration Actually Goes",{"type":42,"tag":43,"props":1751,"children":1752},{},[1753],{"type":47,"value":1754},"The fear is that everything stops for months while engineers rebuild what already works. It doesn't have to, and it shouldn't.",{"type":42,"tag":240,"props":1756,"children":1757},{},[1758,1768,1778],{"type":42,"tag":244,"props":1759,"children":1760},{},[1761,1766],{"type":42,"tag":148,"props":1762,"children":1763},{},[1764],{"type":47,"value":1765},"You keep what you learned.",{"type":47,"value":1767}," Every screen and workflow you built on the platform is a finished spec. The hard part, knowing what to build, is done. That's why a rebuild of a validated app is usually faster than the original guesswork.",{"type":42,"tag":244,"props":1769,"children":1770},{},[1771,1776],{"type":42,"tag":148,"props":1772,"children":1773},{},[1774],{"type":47,"value":1775},"You run both in parallel.",{"type":47,"value":1777}," The platform app stays live and serving users while the real product comes together. Nothing goes dark.",{"type":42,"tag":244,"props":1779,"children":1780},{},[1781,1786],{"type":42,"tag":148,"props":1782,"children":1783},{},[1784],{"type":47,"value":1785},"You migrate the load-bearing pieces first.",{"type":47,"value":1787}," Start with whatever is straining hardest: the slow workflow, the integration you needed, the part that keeps breaking. Get the highest-pain piece onto solid ground first, then move the rest.",{"type":42,"tag":43,"props":1789,"children":1790},{},[1791,1793,1799,1801,1806],{"type":47,"value":1792},"For a broader look at taking an early product all the way to a mature one, our guide on ",{"type":42,"tag":435,"props":1794,"children":1796},{"href":1795},"/blog/scale-mvp-to-product/",[1797],{"type":47,"value":1798},"scaling from MVP to product",{"type":47,"value":1800}," covers the wider journey; this post is the specific case of doing it ",{"type":42,"tag":316,"props":1802,"children":1803},{},[1804],{"type":47,"value":1805},"from a no-code or AI tool",{"type":47,"value":322},{"type":42,"tag":55,"props":1808,"children":1809},{"id":759},[1810],{"type":47,"value":762},{"type":42,"tag":43,"props":1812,"children":1813},{},[1814,1816,1822],{"type":47,"value":1815},"Migrating off a tool takes two things: understanding what you built and knowing how to rebuild it so it lasts. If your app was built with AI and it's the fragility and breakage pushing you to move, that specific situation is exactly what our ",{"type":42,"tag":435,"props":1817,"children":1819},{"href":1818},"/services/ai-app-rescue/",[1820],{"type":47,"value":1821},"AI App Rescue",{"type":47,"value":1823}," work is for: we keep what works and get it to production without starting over.",{"type":42,"tag":43,"props":1825,"children":1826},{},[1827],{"type":47,"value":1828},"If you're weighing the move, that's a good, honest conversation, and we'll tell you plainly whether you're ready for it or better off staying put a little longer.",{"type":42,"tag":55,"props":1830,"children":1831},{"id":783},[1832],{"type":47,"value":786},{"type":42,"tag":788,"props":1834,"children":1836},{"id":1835},"can-i-turn-my-lovable-or-bubble-app-into-a-real-product",[1837],{"type":47,"value":1594},{"type":42,"tag":43,"props":1839,"children":1840},{},[1841],{"type":47,"value":1595},{"type":42,"tag":788,"props":1843,"children":1845},{"id":1844},"do-i-have-to-rebuild-from-scratch-to-move-off-bubble-or-lovable",[1846],{"type":47,"value":1597},{"type":42,"tag":43,"props":1848,"children":1849},{},[1850],{"type":47,"value":1598},{"type":42,"tag":788,"props":1852,"children":1854},{"id":1853},"why-wont-my-bubble-or-lovable-app-scale",[1855],{"type":47,"value":1600},{"type":42,"tag":43,"props":1857,"children":1858},{},[1859],{"type":47,"value":1601},{"type":42,"tag":788,"props":1861,"children":1863},{"id":1862},"when-should-i-migrate-off-a-no-code-or-ai-tool",[1864],{"type":47,"value":1603},{"type":42,"tag":43,"props":1866,"children":1867},{},[1868],{"type":47,"value":1604},{"type":42,"tag":43,"props":1870,"children":1871},{},[1872],{"type":42,"tag":316,"props":1873,"children":1874},{},[1875,1876,1881,1882,1886,1887,1891],{"type":47,"value":521},{"type":42,"tag":435,"props":1877,"children":1878},{"href":1644},[1879],{"type":47,"value":1880},"No-Code vs Custom: When You Hit the Ceiling",{"type":47,"value":837},{"type":42,"tag":435,"props":1883,"children":1884},{"href":1818},[1885],{"type":47,"value":1821},{"type":47,"value":844},{"type":42,"tag":435,"props":1888,"children":1889},{"href":712},[1890],{"type":47,"value":849},{"type":47,"value":322},{"title":7,"searchDepth":551,"depth":551,"links":1893},[1894,1895,1896,1897,1898,1899,1900,1901,1902],{"id":1628,"depth":551,"text":1631},{"id":1652,"depth":551,"text":1655},{"id":1675,"depth":551,"text":1678},{"id":1691,"depth":551,"text":1694},{"id":1714,"depth":551,"text":1717},{"id":1730,"depth":551,"text":1733},{"id":1746,"depth":551,"text":1749},{"id":759,"depth":551,"text":762},{"id":783,"depth":551,"text":786,"children":1903},[1904,1905,1906,1907],{"id":1835,"depth":863,"text":1594},{"id":1844,"depth":863,"text":1597},{"id":1853,"depth":863,"text":1600},{"id":1862,"depth":863,"text":1603},"content:blog:lovable-bubble-app-to-production.md","blog/lovable-bubble-app-to-production.md","blog/lovable-bubble-app-to-production",{"_path":1912,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":1913,"description":1914,"metaTitle":1915,"topic":1916,"author":12,"authorProfile":13,"coverImg":1917,"coverImgAlt":1918,"toc":16,"published":16,"createdAt":1919,"updatedAt":1919,"keywords":1920,"faq":1927,"body":1940,"_type":563,"_id":2272,"_source":565,"_file":2273,"_stem":2274,"_extension":568},"/blog/hire-developer-to-fix-ai-code","How to Hire a Developer to Fix AI-Generated Code","Your AI-built app is breaking and you need help. Here's how to hire the right developer to fix AI-generated code, the green flags to look for, and the red flags to avoid.","How to Hire a Developer to Fix AI-Generated Code (Without Getting Ripped Off)","Hiring and Team Building","/images/blog/hire-developer-to-fix-ai-code.webp","How to hire the right developer to fix AI-generated code, with the green flags and red flags to watch for.","2026-08-15",[1921,1922,1923,1924,1925,1926],"hire developer to fix ai generated code","fix ai generated code","hire developer to fix my app","fix vibe coded app","hire developer to fix ai app","who can fix ai code",[1928,1931,1934,1937],{"question":1929,"answer":1930},"How do I find a developer to fix my AI-generated app?","Look for someone who fixes and ships production code, not just someone who writes new code. The best signal is that they ask to see your app before quoting anything, talk about keeping what works instead of rewriting everything, and can name the usual AI failure spots (tests, security, database, scaling). Start with a paid assessment, not a blind rebuild quote.",{"question":1932,"answer":1933},"Should I rewrite my AI-built app or just fix it?","Usually fix, not rewrite. AI typically got you a genuine 80%, and throwing that away wastes it. A good developer assesses each part and gives you a keep, fix, or rebuild call, then rebuilds only what truly needs it. Anyone who says 'rewrite everything' before looking is guessing.",{"question":1935,"answer":1936},"How much does it cost to fix AI-generated code?","It depends entirely on what's fragile. A security fix and some hardening is very different from a partial rebuild. That's why a proper assessment comes first: it scopes the real work so you get a real number instead of a guess. It's almost always less than a full rebuild.",{"question":1938,"answer":1939},"What are the red flags when hiring someone to fix AI code?","They quote a price without looking at your app, they immediately push a full rewrite, they can't explain what they'd check or why, or they have no real production experience. Any of those means you're likely to spend more and end up in the same place.",{"type":39,"children":1941,"toc":2259},[1942,1947,1952,1961,1967,1972,1984,1989,1995,2000,2050,2056,2061,2110,2116,2121,2156,2161,2167,2172,2177,2195,2199,2204,2208,2213,2218,2223,2227,2232,2236],{"type":42,"tag":43,"props":1943,"children":1944},{},[1945],{"type":47,"value":1946},"Your app works, mostly. You built it with AI, it got you real users, and now it breaks when you touch it. So you've decided to hire someone to fix it. Good call. But here's the trap: fixing AI-generated code is a different job from writing new code, and the wrong hire can make it worse and more expensive.",{"type":42,"tag":43,"props":1948,"children":1949},{},[1950],{"type":47,"value":1951},"So here's how to hire the right person, from a team that does exactly this work. The goal is simple: get it fixed, without getting ripped off.",{"type":42,"tag":43,"props":1953,"children":1954},{},[1955,1959],{"type":42,"tag":148,"props":1956,"children":1957},{},[1958],{"type":47,"value":617},{"type":47,"value":1960}," hire someone who fixes and ships production code, not just someone who writes fresh code. The best ones ask to see your app before quoting, talk about keeping what works instead of rewriting everything, and can name where AI-built apps usually break. Start with a paid assessment, not a blind rebuild quote.",{"type":42,"tag":55,"props":1962,"children":1964},{"id":1963},"why-fixing-ai-code-is-a-different-skill",[1965],{"type":47,"value":1966},"Why Fixing AI Code Is a Different Skill",{"type":42,"tag":43,"props":1968,"children":1969},{},[1970],{"type":47,"value":1971},"Writing new code from a blank page and fixing someone else's (or an AI's) code are different skills, and plenty of good builders aren't good fixers.",{"type":42,"tag":43,"props":1973,"children":1974},{},[1975,1977,1982],{"type":47,"value":1976},"Fixing AI-generated code specifically needs three things at once: the ability to read and diagnose unfamiliar code quickly, an understanding of ",{"type":42,"tag":316,"props":1978,"children":1979},{},[1980],{"type":47,"value":1981},"how",{"type":47,"value":1983}," AI tends to write things (and where it cuts corners), and real production experience so they know what \"safe to ship\" actually means. Someone who has only ever built greenfield projects can struggle here, because the job is detective work before it's building work.",{"type":42,"tag":43,"props":1985,"children":1986},{},[1987],{"type":47,"value":1988},"That's why \"I can build you an app\" is not the same as \"I can fix this app.\" You want the second person.",{"type":42,"tag":55,"props":1990,"children":1992},{"id":1991},"the-green-flags-what-a-good-fixer-does",[1993],{"type":47,"value":1994},"The Green Flags: What a Good Fixer Does",{"type":42,"tag":43,"props":1996,"children":1997},{},[1998],{"type":47,"value":1999},"When you talk to someone, these are the signs you've found the right person:",{"type":42,"tag":240,"props":2001,"children":2002},{},[2003,2013,2023,2040],{"type":42,"tag":244,"props":2004,"children":2005},{},[2006,2011],{"type":42,"tag":148,"props":2007,"children":2008},{},[2009],{"type":47,"value":2010},"They ask to see it first.",{"type":47,"value":2012}," A good developer wants to look at your actual code and app before saying anything about price or timeline. Diagnosis comes before the quote.",{"type":42,"tag":244,"props":2014,"children":2015},{},[2016,2021],{"type":42,"tag":148,"props":2017,"children":2018},{},[2019],{"type":47,"value":2020},"They talk about keeping what works.",{"type":47,"value":2022}," They're looking to fix and harden what's fragile, not to throw away the 80% the AI got right. A rescue, not a teardown.",{"type":42,"tag":244,"props":2024,"children":2025},{},[2026,2031,2033,2038],{"type":42,"tag":148,"props":2027,"children":2028},{},[2029],{"type":47,"value":2030},"They know where AI-built apps break.",{"type":47,"value":2032}," They can quickly name the usual suspects: missing tests, security gaps, database performance, scaling, the code breaking whenever you add a feature. If they know the pattern, they've seen it before. (Here's ",{"type":42,"tag":435,"props":2034,"children":2035},{"href":712},[2036],{"type":47,"value":2037},"where AI-built apps actually break",{"type":47,"value":2039}," so you can check.)",{"type":42,"tag":244,"props":2041,"children":2042},{},[2043,2048],{"type":42,"tag":148,"props":2044,"children":2045},{},[2046],{"type":47,"value":2047},"They explain things plainly.",{"type":47,"value":2049}," They can tell you what's wrong and what they'd do in language you understand, without hiding behind jargon. Clarity is a sign they actually know.",{"type":42,"tag":55,"props":2051,"children":2053},{"id":2052},"the-red-flags-when-to-walk-away",[2054],{"type":47,"value":2055},"The Red Flags: When to Walk Away",{"type":42,"tag":43,"props":2057,"children":2058},{},[2059],{"type":47,"value":2060},"Just as important, here's when to keep looking:",{"type":42,"tag":240,"props":2062,"children":2063},{},[2064,2074,2084,2100],{"type":42,"tag":244,"props":2065,"children":2066},{},[2067,2072],{"type":42,"tag":148,"props":2068,"children":2069},{},[2070],{"type":47,"value":2071},"They quote a price without looking.",{"type":47,"value":2073}," A number before they've seen your app is a guess, and usually a high one to cover themselves.",{"type":42,"tag":244,"props":2075,"children":2076},{},[2077,2082],{"type":42,"tag":148,"props":2078,"children":2079},{},[2080],{"type":47,"value":2081},"They immediately say \"rewrite everything.\"",{"type":47,"value":2083}," Sometimes a rebuild is right, but leading with it before diagnosis usually means it's the only tool they have. It's also the most expensive option, conveniently.",{"type":42,"tag":244,"props":2085,"children":2086},{},[2087,2092,2094,2098],{"type":42,"tag":148,"props":2088,"children":2089},{},[2090],{"type":47,"value":2091},"They can't explain what they'd check.",{"type":47,"value":2093}," If they can't tell you ",{"type":42,"tag":316,"props":2095,"children":2096},{},[2097],{"type":47,"value":1981},{"type":47,"value":2099}," they'd figure out what's wrong, they probably don't know.",{"type":42,"tag":244,"props":2101,"children":2102},{},[2103,2108],{"type":42,"tag":148,"props":2104,"children":2105},{},[2106],{"type":47,"value":2107},"They have no real production experience.",{"type":47,"value":2109}," Someone who's never run software with real users won't know what breaks in production, which is the whole problem you're trying to solve.",{"type":42,"tag":55,"props":2111,"children":2113},{"id":2112},"freelancer-agency-or-dedicated-developer",[2114],{"type":47,"value":2115},"Freelancer, Agency, or Dedicated Developer?",{"type":42,"tag":43,"props":2117,"children":2118},{},[2119],{"type":47,"value":2120},"For this specific job, the honest trade-offs:",{"type":42,"tag":43,"props":2122,"children":2123},{},[2124,2126,2131,2133,2138,2140,2146,2148,2154],{"type":47,"value":2125},"A ",{"type":42,"tag":148,"props":2127,"children":2128},{},[2129],{"type":47,"value":2130},"freelancer",{"type":47,"value":2132}," can be great and affordable if you find one who genuinely fixes production code, but many stop at a working demo, which is how you got here. An ",{"type":42,"tag":148,"props":2134,"children":2135},{},[2136],{"type":47,"value":2137},"agency or dedicated team",{"type":47,"value":2139}," costs more but brings backup: if your fixer hits a wall on security or infrastructure, they have colleagues to pull in, and they've usually seen your exact problem before. We wrote a fuller ",{"type":42,"tag":435,"props":2141,"children":2143},{"href":2142},"/blog/freelancer-vs-agency-software-development/",[2144],{"type":47,"value":2145},"freelancer vs agency comparison",{"type":47,"value":2147}," if you want to weigh it, and our general guide on ",{"type":42,"tag":435,"props":2149,"children":2151},{"href":2150},"/blog/how-to-vet-a-developer/",[2152],{"type":47,"value":2153},"how to vet a developer",{"type":47,"value":2155}," applies here too.",{"type":42,"tag":43,"props":2157,"children":2158},{},[2159],{"type":47,"value":2160},"Whichever route you pick, the vetting above is what matters more than the label.",{"type":42,"tag":55,"props":2162,"children":2164},{"id":2163},"start-with-an-assessment-not-a-rebuild-quote",[2165],{"type":47,"value":2166},"Start With an Assessment, Not a Rebuild Quote",{"type":42,"tag":43,"props":2168,"children":2169},{},[2170],{"type":47,"value":2171},"The single best first step is a paid assessment, not a blind quote for a rebuild.",{"type":42,"tag":43,"props":2173,"children":2174},{},[2175],{"type":47,"value":2176},"A proper assessment means someone looks at your code, your architecture, and how it behaves, then tells you honestly what's solid, what's fragile, and what's a real risk, with a keep, fix, or rebuild call for each part. That gives you a real scope and a real number, instead of a guess. And if the fix turns out to be small, or you don't need much at all, a trustworthy person will tell you that too.",{"type":42,"tag":43,"props":2178,"children":2179},{},[2180,2182,2186,2188,2193],{"type":47,"value":2181},"This is exactly how our ",{"type":42,"tag":435,"props":2183,"children":2184},{"href":1818},[2185],{"type":47,"value":1821},{"type":47,"value":2187}," work starts. We built AI into our own product and we fix AI-built apps for others, so we can tell you quickly whether it's a quick fix or a real project, and price it straight. If you're still earlier and deciding whether to fix it or hand the whole thing to a team, our honest take on ",{"type":42,"tag":435,"props":2189,"children":2190},{"href":1184},[2191],{"type":47,"value":2192},"building with AI versus hiring",{"type":47,"value":2194}," covers that.",{"type":42,"tag":55,"props":2196,"children":2197},{"id":783},[2198],{"type":47,"value":786},{"type":42,"tag":788,"props":2200,"children":2202},{"id":2201},"how-do-i-find-a-developer-to-fix-my-ai-generated-app",[2203],{"type":47,"value":1929},{"type":42,"tag":43,"props":2205,"children":2206},{},[2207],{"type":47,"value":1930},{"type":42,"tag":788,"props":2209,"children":2211},{"id":2210},"should-i-rewrite-my-ai-built-app-or-just-fix-it",[2212],{"type":47,"value":1932},{"type":42,"tag":43,"props":2214,"children":2215},{},[2216],{"type":47,"value":2217},"Usually fix, not rewrite. AI typically got you a genuine 80%, and throwing that away wastes it. A good developer assesses each part and gives you a keep, fix, or rebuild call, then rebuilds only what truly needs it. Anyone who says \"rewrite everything\" before looking is guessing.",{"type":42,"tag":788,"props":2219,"children":2221},{"id":2220},"how-much-does-it-cost-to-fix-ai-generated-code",[2222],{"type":47,"value":1935},{"type":42,"tag":43,"props":2224,"children":2225},{},[2226],{"type":47,"value":1936},{"type":42,"tag":788,"props":2228,"children":2230},{"id":2229},"what-are-the-red-flags-when-hiring-someone-to-fix-ai-code",[2231],{"type":47,"value":1938},{"type":42,"tag":43,"props":2233,"children":2234},{},[2235],{"type":47,"value":1939},{"type":42,"tag":43,"props":2237,"children":2238},{},[2239],{"type":42,"tag":316,"props":2240,"children":2241},{},[2242,2243,2247,2248,2252,2253,2258],{"type":47,"value":521},{"type":42,"tag":435,"props":2244,"children":2245},{"href":1818},[2246],{"type":47,"value":1821},{"type":47,"value":837},{"type":42,"tag":435,"props":2249,"children":2250},{"href":712},[2251],{"type":47,"value":849},{"type":47,"value":844},{"type":42,"tag":435,"props":2254,"children":2255},{"href":2150},[2256],{"type":47,"value":2257},"How to Vet a Developer",{"type":47,"value":322},{"title":7,"searchDepth":551,"depth":551,"links":2260},[2261,2262,2263,2264,2265,2266],{"id":1963,"depth":551,"text":1966},{"id":1991,"depth":551,"text":1994},{"id":2052,"depth":551,"text":2055},{"id":2112,"depth":551,"text":2115},{"id":2163,"depth":551,"text":2166},{"id":783,"depth":551,"text":786,"children":2267},[2268,2269,2270,2271],{"id":2201,"depth":863,"text":1929},{"id":2210,"depth":863,"text":1932},{"id":2220,"depth":863,"text":1935},{"id":2229,"depth":863,"text":1938},"content:blog:hire-developer-to-fix-ai-code.md","blog/hire-developer-to-fix-ai-code.md","blog/hire-developer-to-fix-ai-code",{"_path":571,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":572,"description":573,"metaTitle":574,"topic":11,"author":12,"authorProfile":13,"coverImg":575,"coverImgAlt":576,"toc":16,"published":16,"createdAt":577,"updatedAt":577,"keywords":2276,"faq":2277,"body":2282,"_type":563,"_id":867,"_source":565,"_file":868,"_stem":869,"_extension":568},[579,580,581,582,583,584],[2278,2279,2280,2281],{"question":587,"answer":588},{"question":590,"answer":591},{"question":593,"answer":594},{"question":596,"answer":597},{"type":39,"children":2283,"toc":2477},[2284,2288,2292,2300,2304,2308,2312,2316,2320,2329,2333,2352,2356,2360,2364,2373,2377,2381,2390,2394,2398,2402,2406,2410,2419,2423,2427,2431,2435,2439,2443,2447,2451,2455],{"type":42,"tag":43,"props":2285,"children":2286},{},[2287],{"type":47,"value":604},{"type":42,"tag":43,"props":2289,"children":2290},{},[2291],{"type":47,"value":609},{"type":42,"tag":43,"props":2293,"children":2294},{},[2295,2299],{"type":42,"tag":148,"props":2296,"children":2297},{},[2298],{"type":47,"value":617},{"type":47,"value":619},{"type":42,"tag":55,"props":2301,"children":2302},{"id":622},[2303],{"type":47,"value":625},{"type":42,"tag":43,"props":2305,"children":2306},{},[2307],{"type":47,"value":630},{"type":42,"tag":43,"props":2309,"children":2310},{},[2311],{"type":47,"value":635},{"type":42,"tag":55,"props":2313,"children":2314},{"id":638},[2315],{"type":47,"value":641},{"type":42,"tag":43,"props":2317,"children":2318},{},[2319],{"type":47,"value":646},{"type":42,"tag":43,"props":2321,"children":2322},{},[2323,2324,2328],{"type":47,"value":651},{"type":42,"tag":316,"props":2325,"children":2326},{},[2327],{"type":47,"value":656},{"type":47,"value":658},{"type":42,"tag":55,"props":2330,"children":2331},{"id":661},[2332],{"type":47,"value":664},{"type":42,"tag":43,"props":2334,"children":2335},{},[2336,2337,2341,2342,2346,2347,2351],{"type":47,"value":669},{"type":42,"tag":316,"props":2338,"children":2339},{},[2340],{"type":47,"value":674},{"type":47,"value":676},{"type":42,"tag":316,"props":2343,"children":2344},{},[2345],{"type":47,"value":674},{"type":47,"value":682},{"type":42,"tag":316,"props":2348,"children":2349},{},[2350],{"type":47,"value":674},{"type":47,"value":688},{"type":42,"tag":43,"props":2353,"children":2354},{},[2355],{"type":47,"value":693},{"type":42,"tag":55,"props":2357,"children":2358},{"id":696},[2359],{"type":47,"value":699},{"type":42,"tag":43,"props":2361,"children":2362},{},[2363],{"type":47,"value":704},{"type":42,"tag":43,"props":2365,"children":2366},{},[2367,2368,2372],{"type":47,"value":709},{"type":42,"tag":435,"props":2369,"children":2370},{"href":712},[2371],{"type":47,"value":715},{"type":47,"value":322},{"type":42,"tag":55,"props":2374,"children":2375},{"id":719},[2376],{"type":47,"value":722},{"type":42,"tag":43,"props":2378,"children":2379},{},[2380],{"type":47,"value":727},{"type":42,"tag":43,"props":2382,"children":2383},{},[2384,2385,2389],{"type":47,"value":732},{"type":42,"tag":435,"props":2386,"children":2387},{"href":735},[2388],{"type":47,"value":738},{"type":47,"value":740},{"type":42,"tag":55,"props":2391,"children":2392},{"id":743},[2393],{"type":47,"value":746},{"type":42,"tag":43,"props":2395,"children":2396},{},[2397],{"type":47,"value":751},{"type":42,"tag":43,"props":2399,"children":2400},{},[2401],{"type":47,"value":756},{"type":42,"tag":55,"props":2403,"children":2404},{"id":759},[2405],{"type":47,"value":762},{"type":42,"tag":43,"props":2407,"children":2408},{},[2409],{"type":47,"value":767},{"type":42,"tag":43,"props":2411,"children":2412},{},[2413,2414,2418],{"type":47,"value":772},{"type":42,"tag":435,"props":2415,"children":2416},{"href":775},[2417],{"type":47,"value":778},{"type":47,"value":780},{"type":42,"tag":55,"props":2420,"children":2421},{"id":783},[2422],{"type":47,"value":786},{"type":42,"tag":788,"props":2424,"children":2425},{"id":790},[2426],{"type":47,"value":587},{"type":42,"tag":43,"props":2428,"children":2429},{},[2430],{"type":47,"value":588},{"type":42,"tag":788,"props":2432,"children":2433},{"id":799},[2434],{"type":47,"value":590},{"type":42,"tag":43,"props":2436,"children":2437},{},[2438],{"type":47,"value":591},{"type":42,"tag":788,"props":2440,"children":2441},{"id":808},[2442],{"type":47,"value":593},{"type":42,"tag":43,"props":2444,"children":2445},{},[2446],{"type":47,"value":594},{"type":42,"tag":788,"props":2448,"children":2449},{"id":817},[2450],{"type":47,"value":596},{"type":42,"tag":43,"props":2452,"children":2453},{},[2454],{"type":47,"value":597},{"type":42,"tag":43,"props":2456,"children":2457},{},[2458],{"type":42,"tag":316,"props":2459,"children":2460},{},[2461,2462,2466,2467,2471,2472,2476],{"type":47,"value":521},{"type":42,"tag":435,"props":2463,"children":2464},{"href":775},[2465],{"type":47,"value":835},{"type":47,"value":837},{"type":42,"tag":435,"props":2468,"children":2469},{"href":735},[2470],{"type":47,"value":842},{"type":47,"value":844},{"type":42,"tag":435,"props":2473,"children":2474},{"href":712},[2475],{"type":47,"value":849},{"type":47,"value":322},{"title":7,"searchDepth":551,"depth":551,"links":2478},[2479,2480,2481,2482,2483,2484,2485,2486],{"id":622,"depth":551,"text":625},{"id":638,"depth":551,"text":641},{"id":661,"depth":551,"text":664},{"id":696,"depth":551,"text":699},{"id":719,"depth":551,"text":722},{"id":743,"depth":551,"text":746},{"id":759,"depth":551,"text":762},{"id":783,"depth":551,"text":786,"children":2487},[2488,2489,2490,2491],{"id":790,"depth":863,"text":587},{"id":799,"depth":863,"text":590},{"id":808,"depth":863,"text":593},{"id":817,"depth":863,"text":596},1788240160043]