Clarity
We say what a system will and will not do before we build it. Scope, risks and trade-offs are written down, not implied.
[01]About Built for the AI era
We design and ship the software that thinks alongside your team.
Four brings product thinking, applied AI and hard-nosed engineering into one small, senior studio. We take ideas from a whiteboard sketch to a system your organisation relies on every day, and we stay accountable for how it behaves in production.
- Systems shipped to production
- 120+
- Years of engineering practice
- 16
- Disciplines, one team
- 4
[02]Selected work Illustrative projects
Systems that earn their place in the workflow.
01 / 04
- 01
Signal
2026
Agentic support desk
- 02
Tideline
2025
Realtime forecasting
- 03
Ledger
2025
Document intelligence
- 04
Relay
2024
Voice-first operations
[03]Disciplines Four disciplines, one team
Deliberately focused.
We work where complexity is high and reliability is non-negotiable, applying the same engineering discipline to every model we put into production.
No. Discipline Projects*
-
Autonomous workflows with guardrails, evaluations and a human in the loop wherever it matters.
-
Retrieval, structured output and tool use wired into the systems you already run.
-
Fast, accessible web products with the boring parts done properly: auth, data, deploys.
-
Clean, observable data flows that make models useful and reporting trustworthy.
*Illustrative figures for a fictional studio.
[04]Principles How we work
Four principles?
Three. We keep it honest.
- 01Clarity
- 02Craft
- 03Candour
Craft
Small details decide whether software is trusted. We test, measure and polish until the edges hold up under real use.
Candour
If a model is the wrong tool, we tell you. Honest advice early beats an impressive demo that fails in month three.