How we use AI
Built by a senior team.
Accelerated by AI.
Our senior team works with Claude Code, Codex and models from Anthropic and OpenAI. The decisions, and the sign-off on every result, stay with people.
Want AI inside your own product? See AI integration
What changes for you
Senior standards, with less waiting between steps.
Working versions sooner
Prototypes and running builds early, not after a long spec.
More options before you commit
Several directions explored while a change is still cheap.
Senior time on the hard parts
Routine work goes to AI, so senior hours go into decisions.
AI at every step
AI drafts. A senior decides.
Understand
AI sorts the interviews
A senior picks the problem
Explore
AI drafts the variants
A senior picks one direction
Build
Agents write routine code
A senior owns the architecture
Validate
AI writes the tests
A senior decides what must pass
Release
AI drafts the release notes
A senior decides when it ships
Development, design, marketing
Where AI saves real effort
People decide
Six decisions that stay with people.
Use the arrow keys to switch between the six decisions.
A senior lead sets the scope of each release with you.
Seen in the work
An agent platform MVP was scoped to fit two months, and delivered in two.
AI agent platform MVPTools and rules
Named tools. Clear rules.
- 01
Claude Code and Codex
Coding agents from Anthropic and OpenAI, run by our senior engineers.
- 02
Services and access
External services and data access are agreed upfront.
- 03
Not used for training
Your data is excluded from AI provider training.
- 04
Your infrastructure
Hosting within your own infrastructure is supported.
- 05
Yours from day one
You own the code and design files, with access from day one.
See where AI fits
in your project.
A short call with the senior team who would build it: where we would use AI, where we would not, and how your data is handled.
Questions about how we use AI
Claude Code by Anthropic and Codex by OpenAI, and models from both companies.
It makes your budget go further. We agree a fixed price for a clear scope before development starts, and the time AI saves goes into key scenarios, quality and edge cases.
AI output is a draft until a senior accepts it. Every change is reviewed by a second senior engineer, and you accept each milestone against the examples you agreed to.
External services and data access are agreed upfront, before the work starts. Your data is excluded from AI provider training, and hosting within your own infrastructure is supported.
You approve the scope, set the business priorities and accept each milestone. A decision owner and someone who knows the workflow are usually enough on your side. See how we work for each stage.
The people who built it. Bugs within the delivered scope are fixed at no extra cost for 30 days after handover, and ongoing support and improvements can be arranged.
No. This page is about how our team works. If you want AI inside your product, such as search over your documents or an agent that answers customers, see AI integration.






