What we can show you.

Past work is under confidentiality agreements, so you will not find logos or case studies here. This is what we can show you instead.

Experience

Production conversational AI for a US insurtech.
Running in production, under confidentiality terms.
Enterprise systems delivery.
Delivery inside large organisations, with their reviews, security and release processes.

Working software says more than a logo wall. These run on this site.

Seven questions that map your AI system to its obligations and deadlines. Runs entirely in your browser.

Approve or decline a payout in the illustrative claims agent on the home page.

A private spam check, email confirmation and scam signals, built from scratch.

Designed, rendered and engineered in-house, fast and accessible.

A typical shape for one automated workflow. Every project is scoped in discovery, so treat this as a guide, not a promise.

Discovery
Week 1
We map one workflow end to end: inputs, systems, data, risks and what good looks like. You get a written plan and a quote.
First working version
Weeks 2–3
The core path runs on real examples. You see it in a live demo, not a slide.
Hardening
Weeks 4–5
Edge cases, human-review thresholds, logging and an evaluation set built from your data.
Launch
Week 6
Monitoring, runbook and handover notes. Your team knows how it works and how to change it.
Improve or hand over
After
We keep improving it with you, or hand it over cleanly. Either way you own the code and the accounts.

Start with a one-week discovery. You get a written plan and a quote, and you can decide from there with no commitment beyond it.