AI product development
We build MVPs, SaaS platforms, mobile apps and AI features inside existing products, to production standard.
- Specification sprint: £4,500 for one week, £8,500 for two
- You keep the specification
Apps, SaaS platforms and MVPs we build
AI product development is building software products, and the AI features inside them, to production standard: a written specification first, a senior engineer's review of every change and automated tests that must pass before anything ships.
Quantum Craft does AI product development for funded startups and UK companies. A senior engineer reviews every change against a written specification. A single-workflow product built to production standard typically costs £35,000 to £80,000 over 6 to 12 weeks; larger platforms such as BlueWave and Karu each had a core build of about 14 weeks.
Last reviewed
Most of our work takes one of four shapes.
The four shapes of work, each beside what its card links to: MVPs and SaaS platforms to BlueWave and Karu, mobile apps to NForge, AI features inside products to a B2B logistics marketplace and taking over existing code to vibe code rescue. The numbers are the cards' numbers.
- 1. MVPs and SaaS platforms
- 2. Mobile apps
- 3. AI features inside products
- 4. Taking over existing code
MVPs and SaaS platforms
A first version with one workflow done properly, or a multi-tenant platform with roles, billing and an admin console behind it, like BlueWave. We keep customers apart in the data layer, where a mistake fails loudly, instead of trusting the interface to hide one account from another. Karu is an example.
Mobile apps
Native iOS and Android from one Flutter codebase, as in NForge's coach app, and offline-first field apps that keep working in a basement plant room, then sync when the engineer gets signal back. We have written up the offline architecture we use.
AI features inside products
Assistants that answer from your own records, document extraction, agents that act through a small set of tools, and generation pipelines whose output is checked before anyone sees it. A B2B logistics marketplace we built runs five AI features in production. Before you commission an agent, check the job needs one: a plain workflow is often cheaper to run and easier to test.
Taking over existing code
Code another team wrote, including prototypes generated with AI app builders such as Lovable, Bolt.new or Replit. We read it before we quote, keep what is sound and replace what is not. For apps built mostly by prompting, vibe code rescue starts with an AI-built app audit at a fixed £3,500.
How much does an MVP cost in the UK, and how long does it take?
A focused MVP typically costs £35,000 to £80,000 over 6 to 12 weeks. Every build starts with a specification sprint: the written specification, the architecture, an estimate and the risks, for £4,500 over one week on a single workflow or £8,500 over two weeks on a platform. The document is yours, and the build is quoted from it. The table gives the shapes we have built.
Every published band on one linear scale from zero, numbered in the key. Highlighted: the single-workflow MVP, the band this page leads with.
- 1. Single-workflow MVP, £35,000 to £80,000
- 2. Practice platform, £90,000 to £140,000 for this phase
- 3. Subscription learning platform, £110,000 to £175,000
- 4. Two-sided marketplace, £125,000 to £170,000
- 5. Field-service compliance platform, £150,000 to £250,000 to first release
- 6. Platform with native apps and payments, £260,000 to £340,000
- 7. Data platform programme, £350,000 to £550,000 across the programme
Indicative bands by shape of build
Shape
Single-workflow MVP
- Example engagement and its scope
- One workflow built to production standard. The bottom of the band has one user type and no integrations; the top adds payments, an admin surface and a second audience
- Time
- 6 to 12 weeks
- Indicative band
- £35,000 to £80,000
Practice platform
- Example engagement and its scope
- Karu: scheduling, records, notes and video
- Time
- About 14 weeks of core build
- Indicative band
- £90,000 to £140,000 for this phase
Subscription learning platform
- Example engagement and its scope
- Treble: student web app, authoring studio with playable notation, marking engine and Stripe billing
- Time
- Four phases
- Indicative band
- £110,000 to £175,000
Two-sided marketplace
- Example engagement and its scope
- A B2B logistics marketplace: the first transactable phase
- Time
- Phase one of a multi-year programme
- Indicative band
- £125,000 to £170,000
Field-service compliance platform
- Example engagement and its scope
- BlueWave: web app, offline mobile app, scheduling and billing
- Time
- About 14 weeks of core build to first release
- Indicative band
- £150,000 to £250,000 to first release
Platform with native apps and payments
- Example engagement and its scope
- NForge: web and native mobile apps, Stripe Connect payments and live video
- Time
- 30 to 40 weeks
- Indicative band
- £260,000 to £340,000
Data platform programme
- Example engagement and its scope
- TIYO: a crawl engine, an event store, analytics and several products
- Time
- Two to three years
- Indicative band
- £350,000 to £550,000 across the programme
All bands exclude costs paid to third parties such as hosting and infrastructure, at London rates as of mid-2026. They are bands for the shape, not an invoice: your figure comes from the specification. Once the product is live, plan for 20 to 30 per cent of the build cost a year to run it and keep changing it.
What the money and the time buy, in order: a specification sprint of one week for £4,500 or two weeks for £8,500; then the build, which takes 6 to 12 weeks for a focused MVP; then, once the product is live, 20 to 30 per cent of the build cost a year to run it and keep changing it. Highlighted station: the specification sprint, where every build starts.
Tell us what the product has to do and by when, and a senior engineer will say what a first release would take.
Request a free 30-minute callAdding AI features and AI agents to your product
A model that impresses in a demo tends to fail in production in familiar ways. It states things that are not true, and it costs more each month than anyone planned. Then the provider retires the version it was tested on. We build the controls in from the first release. Each one below comes from a client build.
A product feature: one model step inside its six controls, numbered as the cards. A case that fails and cannot be repaired goes to a person, as the card on rules enforced in code describes. Highlighted: control 2, rules enforced in code.
- 1. Evaluation before launch
- 2. Rules enforced in code
- 3. Agents act through narrow tools
- 4. Cost you can see per call
- 5. More than one model
- 6. Personal data stays where it should
- Iron mark: a person takes the case that cannot be repaired.
Evaluation before launch
An AI feature gets a fixed set of test scenarios, run again whenever the prompt or the model changes. NForge's programme generator has an evaluation harness that runs the real pipeline over synthetic and replayed intakes, and plants known defects to check that its verifier catches them.
Rules enforced in code
Anything that must be true is checked by code on the server. NForge computes calories from macros instead of trusting the model, and around thirty validation gates run before a coach sees a programme; one that cannot be repaired goes to a person. On a B2B logistics marketplace we built, the assistant cannot present a warehouse unless a real tool call returned it.
Agents act through narrow tools
An agent gets named tools with a declared effect, rather than open access to a database. The same marketplace exposes its operations to AI agents through an MCP server, and every tool is classified as read, create, mutate or destructive.
Cost you can see per call
We log what every AI call costs, by call and by model. NForge's public trial has rate caps, a daily budget breaker and a kill switch, because a free demo running on a paid model is an open tab.
More than one model
The main model is chosen in a bake-off on results and cost, with a fallback ready. The logistics marketplace runs several providers behind weighted A/B tests. When a price rises or a model is retired, we re-run the evaluation instead of rewriting the feature.
Personal data stays where it should
For AI over sensitive personal data we start from Karu's rules: inference in the UK or EU, terms that bar training on your customers' data, a data protection impact assessment first, and features that ship switched off. Karu's clinical drafting layer is designed to those rules and has not shipped yet.
Where each of the six controls comes from: NForge, B2B logistics marketplace and Karu. A line joins a client build to each control whose card names it, and the numbers are the cards' numbers.
Will the code pass technical due diligence?
It will if the evidence is in the repository. Investors now ask how code was checked and who reviewed it, and "we were careful" is not an answer anyone can verify. This is what we give them to check.
Tests that fail without the change
We write each test to fail if the code it covers breaks, so a green build means something. Automated tests run in every build of BlueWave and NForge alike, and a failing build does not ship.
Decisions with their reasons
Architecture decisions go on record with their reasoning, so a reviewer can see why the system is shaped the way it is. BlueWave's decisions were on record from its discovery phase.
Senior review before every merge
A senior engineer reviews every change before it is merged, and checks it against the specification.
Code that is yours
Ownership passes to you on full payment, and the repository can sit in your own organisation from the first commit, so a reviewer sees the whole history. Documentation is written alongside the code, so nothing has to be pieced together for the review.
Numbers from our builds
Six platforms, three live, two in build and one delivered. Each figure below is published in the platform's own case study.
- 5 yearsof compliance records, with an audit trail on each
- 30or so checks on every AI-generated programme before a coach sees it
- 0logins for a therapist's clients to lose, because they never get one
- 5AI features in production, with an evaluation rig
- 22retailers tracked across 13 countries
- A 12-week template
The phone shows week 3 of a programme.
- A day, listed
Thursday's sessions and meals sit under the week strip.
- Revenue at a glance
A revenue card with a bar chart from February to July.
- The next payout
A strip says what is paying out, and when: tomorrow.
What the platforms we build run on
These are the technologies under the platforms we have built recently. We choose per product, and we would rather use a dull tool we know well than learn a fashionable one on your budget.
The stack behind our recent builds
- Back end
- .NET 8 and FastEndpoints, with CQRS and event sourcing where the history matters
- Data
- PostgreSQL, Elasticsearch, Redis and Bigtable
- Web
- Next.js 16, React 19, TypeScript and Tailwind
- Mobile
- Flutter for native iOS and Android, and offline-first web apps on IndexedDB
- Infrastructure
- Google Cloud (Kubernetes in the London region, Pub/Sub and Cloud Storage), AWS and Azure
- Identity, payments, video
- FusionAuth, Stripe and Stripe Connect with Bacs Direct Debit, and LiveKit
- AI in products
- Model APIs from more than one provider, evaluation rigs and MCP servers
Platforms we build on
- Google Cloud
- Amazon Web Services
- Microsoft Azure
- Stripe
- App Store
- Google Play
- Mapbox
Who this is for
Best for
- Funded founders who need a first release that survives real users and technical due diligence
- UK companies adding AI features, such as assistants, search or document extraction, to a product they already run
- Products that take payments or hold regulated data, where the hard parts need senior hands
- Teams that want the specification, the code and the accounts in their own name from the first commit
Not for
- Testing an idea cheaply. A prototyping tool may be the better purchase for that
- Building without a written specification. We write one first, and it is not an optional step
- A large team from the first week: we are a small senior team
Not sure what to build yet?
Start with AI consultancy, not a build. The advice comes from the engineers who would do the building, and sometimes it is to buy an existing product, or to wait. That is cheaper to hear before you spend money on code.
Where to start, from the lists and the paragraph above: testing an idea cheaply points to a prototyping tool, deciding what to build points to AI consultancy, and a first release that has to survive real users and technical due diligence starts with a specification sprint, and then a build. Highlighted station: the specification sprint.
- 1. Testing an idea cheaply
- 2. Deciding what to build
- 3. A first release that survives real users and technical due diligence
Questions buyers ask before a build
What does a software and AI development agency build?
Mostly ordinary software with AI inside it. For us that means MVPs, SaaS platforms and mobile apps, and AI features in products you already run: assistants that answer from your own records, document extraction, search, and agents that act through a few narrow tools. A B2B logistics marketplace we built runs five AI features in production, and NForge's programme generator runs around thirty checks before a coach sees a plan.
How much does MVP development cost in the UK?
A focused MVP, one workflow built to production standard, typically costs £35,000 to £80,000, over 6 to 12 weeks. The bottom of the band has one user type and no integrations; the top adds payments, an admin surface and a second audience. The specification sprint that comes first costs £4,500 or £8,500.
How long does it take to build an AI MVP?
A focused MVP takes 6 to 12 weeks, after a specification sprint of one to two weeks. BlueWave and Karu, both bigger than an MVP, each had a core build of about 14 weeks. An AI feature adds time for evaluation, which we plan from the first week instead of squeezing it in before launch.
8 more questions
Can you take an AI-enabled SaaS product from MVP to production?
Yes. We read what exists first: the code, the data model, the hosting and any AI features. We write down what stands between the MVP and real customers, such as separated customer data, tests on the paths that take money, restore-tested backups, evaluation and cost limits for AI features, and a handover. If an AI app builder made the MVP, an AI-built app audit comes first, at a fixed £3,500.
Can you add AI features to our existing product?
Often that is a better first step than a new product. We read your codebase first, then build the feature inside your existing permissions and data rules, with an evaluation set in place before launch. Document extraction and assistants over your own records are the usual starting points.
We already have a codebase. Can you take it over?
Yes. We read it first, write down what we find, put tests around the parts that matter, and then fix what is blocking your next release. The code might come from an agency, an in-house team or an AI tool.
How do you keep our data secure when AI is involved?
Any AI tool we use works on your code only through business accounts whose terms bar training on it. It never holds production credentials, and client personal data stays out of it. Inside your product, AI features use providers whose terms bar training on your data, with processing in the UK or EU where the data needs it.
Why hasn't AI made custom software cheaper?
Because a product that has to run for years still needs the specification, the architecture, the build, the review and the testing, and none of it can be skipped. Our bands reflect a senior team responsible for all of it.
Can the code sit in our own repository from the first commit?
Yes. The repository can be yours from the first commit, and the code we write for your project is yours on full payment, as our terms set out. Our own tools and methods stay ours, and the code does not need them to run.
What happens after launch?
Plan for 20 to 30 per cent of the build cost a year to run the product and keep changing it. Ongoing work usually runs as a retainer, priced in writing before it starts. If you would rather take the product in-house, the handover is written down and the code is documented for your team.
Which AI model will our product use?
Whichever wins an evaluation on your own data, weighing quality, cost and where the data is allowed to go. You see the evaluation results before we commit to a model.
Tell us what you're building.
Tell us about your product. On a 30-minute call we'll ask who it is for and what it has to do by when, then tell you what a first release would take.
What happens after you get in touch
We reply within one working day
By email, to arrange a time for the call.
A free 30-minute call
With a senior engineer, about what you are building or deciding.
The fee in writing first
If a first step is worth taking, you get its exact fee in writing before any work starts.