THE RECORD

Real engagements, real numbers.

No composites and no hypotheticals. We publish each log as the client signs it off, which takes slightly longer and is worth it.

LOG 01 · OPTIWISER

A product they were buying, rebuilt better in 10 hours.

A five-engineer ML shop replaced a paid SaaS product with an in-house build on Claude, shipped it to clients within a week and kept the capability.

10 hrsTo rebuild the product they were paying for, evals and tests included.
5xFaster software development across the team, our estimate after adoption.
€20kAnnual vendor subscription no longer needed.

LOG 02 · VIRTUALMINDS

From invisible shadow AI to a governed, measurable AI organisation.

A 200-engineer adtech organisation went from no sanctioned tools and invisible AI use to full telemetry, governance and a trained first wave in five months.

200Engineers brought under one visible, governed AI capability.
5 monthsFrom standing start to strategy, budget, handbook, tooling and training.
50Engineers in the first licensed wave, trained weekly as the tools landed.

LOG 03 · COMPLIANCE.ONE

The compliance lawyers went AI-first, with regulated EU data.

A lawyer-led compliance firm went AI-first across product, delivery and sales in about four months, with regulated EU data handled by design and not one engineer hired.

0Engineers hired to take the whole firm AI-first.
~4 monthsFrom start to AI-first across product, delivery and sales.
100%Of AI use routed through one governed, compliant layer.

LOG 04 · ANONYMISED

From knowledge sprawl to a unified knowledge base, compliant by design.

A 30-engineer adtech unit had its institutional knowledge scattered across Confluence, Jira and Git. We built one self-hosted knowledge base with a chat and coding-agent front end, compliant by design, and rolled it out to more than 50 users in weeks.

50+Users across the whole unit, not just the engineers.
1Knowledge base, replacing scattered Confluence, Jira and Git.
15 minReindex cadence, so answers track the current docs.

LOG 05 · ACTIVE AGENT

Agents build the features. Humans still decide what ships.

A 50-engineer adtech platform was losing sprints to tech debt. We built a governed pipeline where cheap, autonomous agents take a ticket to a merge request, and a person signs off everything that reaches the main branch.

~$1To build a feature, down from a developer-day. A review costs cents.
2xDelivery capacity at the same headcount, our estimate with a backlog to pull from.
50%+Cut in lead time, ticket to merged change. Our first measured signal.

LOG 06 · ANONYMISED

A whole commercial engine, run by four people on generative AI.

A high-volume consumer AI platform runs its entire go-to-market engine, generated image and video content for advertising campaigns, automated SEO, lifecycle marketing and sales, on automated generative AI, operated by a four-person team instead of several teams and outside agencies.

4People running the entire GTM engine, content, marketing and conversion.
6-figureAnnual agency, tooling and headcount spend the engine replaces, our estimate.
FixedVideo becomes a fixed cost, not a per-clip bill that climbs with every clip.

LOG 07 · QUINN

Your financial control room, rebuilt as an AI agent.

A fractional CFO built Quinn to run the month-end close: a finance-grade agent that turns days of manual close work into hours. It is in production on real company books, with deterministic data extraction, an authoritative source of record and a human in the loop wherever money or judgement is on the line.

~10 hrsFinance time saved on every month-end close, our estimate: the reconciliation, chasing and report build, automated.
€1k/moFinance time given back each month
1Finance person can own a close that used to take a small team.

LOG 01 · OPTIWISER

A product they were buying, rebuilt better in 10 hours.

A five-engineer ML shop replaced a paid SaaS product with an in-house build on Claude, shipped it to clients within a week and kept the capability.

10 hrsTo rebuild the product they were paying for, evals and tests included.
5xFaster software development across the team, our estimate after adoption.
€20kAnnual vendor subscription no longer needed.

LOG 02 · VIRTUALMINDS

From invisible shadow AI to a governed, measurable AI organisation.

A 200-engineer adtech organisation went from no sanctioned tools and invisible AI use to full telemetry, governance and a trained first wave in five months.

200Engineers brought under one visible, governed AI capability.
5 monthsFrom standing start to strategy, budget, handbook, tooling and training.
50Engineers in the first licensed wave, trained weekly as the tools landed.

LOG 03 · COMPLIANCE.ONE

The compliance lawyers went AI-first, with regulated EU data.

A lawyer-led compliance firm went AI-first across product, delivery and sales in about four months, with regulated EU data handled by design and not one engineer hired.

0Engineers hired to take the whole firm AI-first.
~4 monthsFrom start to AI-first across product, delivery and sales.
100%Of AI use routed through one governed, compliant layer.

FLIGHT HOURS

Companies our people have worked with.

Our crew have built technology for the following companies.

Want a result like these on your books?

The free assessment is how these engagements would start today: five minutes of questions and an instant AI readiness score, then a call with a founder.

Start the assessment Talk to us