Services
Find the right place for AI. Build tools your teams use every day.
We find the use cases worth building, test them on your data and put the best one into production. One senior team, on a fixed fee.
What we do
Strategy and delivery, from one team.
Most companies do the strategy first and then the build. You can also do just one: take the plan and stop there, or bring a use case you’ve already tested and go straight to building.
01 · Strategy
AI Strategy
Discovery Sprint
We find the AI use cases that will make or save you money, test them against your data and systems, and give you a plan you can act on.
- Duration
- 3–4 weeks, fixed fee
- You get
- Ranked use cases, a prototype, a business case and a roadmap
- Right for
- Companies that know AI matters, but not where to start
02 · Build
Custom AI Agents
Production build
We build an AI agent system around your processes, put it into production and hand it over to your team, code included.
- Duration
- 10–12 weeks, fixed fee
- You get
- A live agent system, with its tests, runbooks and source code
- Right for
- Companies with a tested use case, ready to build
AI Strategy
Find where AI changes your costs first.
It takes three to four weeks. At the end you decide whether the build is worth funding.
If the numbers don’t add up, we tell you. A sprint that ends with don’t build this has saved you a lot more than it cost.
- Step 01
Find the use cases that matter
We start with your numbers, not the technology: where costs are highest, where work gets redone, where decisions wait on busy people. Our AI interviews talk to your whole team at once, so we see how work really gets done. You end up with 10 to 12 candidates, each with a value attached.
- Step 02
Check what’s feasible
We score each candidate on how much it’s worth and whether your data and systems can support it today. Most won’t make the cut. It’s far cheaper to drop a weak idea in week two than in month nine.
- Step 03
Build a prototype on your data
We build the strongest candidate on your real data, including the messy cases, since that’s where most AI projects fail. You see it working before you commit any budget.
- Step 04
Make the business case
Expected benefit, cost to build, cost to run and payback, with every assumption written down. Plus a roadmap in order of payback that your finance team can check.
Custom AI Agents
We don’t ship demos. We ship systems.
Getting a demo running takes a week. Getting it to work reliably in a real process takes months, and that’s the part we do. You see working software every week.
Weeks 1–2 · Agree the scope
Before we write any code, we agree the scope, data, integrations and security with your team, and choose the model provider and hosting, including EU hosting if you need it.
Weeks 3–10 · Build and test
We build in weekly steps on your real data. Tests come first, so we can show the agent gets things right, and we tune it to how your teams actually work.
Weeks 11–12 · Go live and hand over
Security sign-off, documentation, runbooks and training for your team. The system runs on your infrastructure, under your accounts.
After launch · Support
The same team stays with you for four weeks after go-live. After that, you run it yourselves or keep us on a monthly retainer.
From you, we need someone who owns the result, the people who run the process, and access to your systems. We handle the rest.
Our platform
A platform underneath every project.
Every project needs the same groundwork, so we built our own tools for it. That way the time you pay for goes into your process, not into setup.
Assessment
AI interviewsRuns anonymous 30-minute AI interviews that adapt to each answer.
Second Brain
Company knowledgeHolds what your company knows, checked by your people and tested with real questions.
Gateway
Model accessSigns in, logs and routes every model call to the provider you approve.
AI Factory
How we buildCoding agents take on tasks and propose changes, and an engineer reviews each one.
You don’t need to buy the platform to work with us. The platform stays ours; the application we build and the data from your project are yours.
How we work
Senior teams that work inside your company.
Three to five people who have built and run AI and data systems at global tech and consulting firms. The people who plan your project are the ones who build it.
- 01
No hand-offs, no juniors
A domain expert, a product lead and engineers who work inside your company build each feature together. Nobody learns on your budget, and nothing gets lost between teams.
- 02
We use AI to build AI
Our AI Factory handles much of the coding, testing and documentation, so the hours you pay for go into your business logic and the hard cases.
- 03
A fixed fee, so the risk is ours
We agree the scope and price up front. If it takes longer than we said, that’s our problem, not a change request.
- 04
You own what we build
The application’s source code, its runbooks and the data from your project stay on your infrastructure.
Who it’s for
Built for companies that run on operations.
Being upfront about this saves us both time.
Mid-sized companies and enterprises, in any industry.
You want to change how you work.
Not just add another tool, or run a pilot to show the board.
You run a real operation.
Orders, claims, tickets, invoices, shipments, cases, and the volume is growing faster than your team.
Someone owns the number.
There’s a line in the P&L this would move, and a person accountable for it.
Your data is messy, but it’s there.
Spreadsheets, old systems, inconsistent naming. That’s normal, and we build for it.
You want to own what we build.
The application’s source code and the data from your project.
You can decide in weeks.
Fast delivery needs fast decisions. Two people who can say yes beat a steering committee.
Getting started
Start where you are.
It depends on how clear you already are about what you need.
Free, no commitment
30-minute call
Bring a problem. We’ll tell you whether AI is the right tool for it, what it would take and roughly what it would cost. If the answer isn’t AI, we’ll say so.
Right for · Still exploring
Book a callFixed fee · 3–4 weeks
Discovery Sprint
You want to find where AI pays off, or you have a few ideas and need to know which one is worth the budget. You get a prototype, a business case and a roadmap, and the plan is yours even if you stop there.
Right for · Before you commit budget
Start with strategyFixed fee · 10–12 weeks
Custom AI Agents build
You already know what you need. We build it, put it into production in your systems and hand over the code and runbooks.
Right for · A tested use case
Start the build
Can’t we build this ourselves?
You can, and some companies should. What an internal team usually lacks is not talent. It is a method for taking one workflow to production reliability, and time away from the existing roadmap. We bring the method, build alongside your people and hand over, so the second workflow can be yours.
We have already prioritised our use cases internally. What would you add?
A test of whether the list survives contact with your data and systems. We score every candidate on value at stake and on feasibility against what you actually have, not what the architecture diagram shows. Most shortlists change. You see which items move up, which drop out and why.
Do we need to clean up our data first?
No. Messy data is the normal case: spreadsheets, legacy systems, inconsistent naming. We check feasibility against the data you actually have and fix only what the first use case needs. A company-wide clean-up before starting usually becomes the reason a programme never starts.
Why did our last pilot never get past proof of concept?
Most likely because nobody defined what “good enough for production” meant. A demo takes a week. Proving that it handles the awkward real cases reliably takes far longer, and that is where programmes quietly die. We agree the acceptance criteria before building. If you have a stalled prototype or agents someone else built, we can start from those.
Do you guarantee the ROI? And does ROI mean headcount?
No guarantee, and be careful with anyone who offers one. The business case shows expected benefit, cost to build, cost to run and payback, with the assumption behind each figure and what has to stay true for it to hold. ROI does not mean headcount by default: the usual goal is to absorb growing volume without adding people at the same pace. What we do commit to is the scope and price of our own work.
Who does the work, and how is it priced?
Three to five senior people: a domain expert, a product lead and forward-deployed engineers. No juniors learning on your budget. The price is a fixed fee for both the Discovery Sprint and the build, agreed up front. If the work takes longer than we said, that is our problem, not a change request.