For regulated and safety-critical organisations
Your team can build on AWS, Azure or Databricks. What it has not done is take an AI system through security, compliance and regulatory sign-off. We bring the platform engineering and the assurance expertise in one senior team, so you deploy with evidence instead of hope.
Built for
COOs, CTOs and CISOs accountable for AI they did not personally build
Product leads stuck between delivery pressure and governance review
IT, security and compliance who need controls they can evidence
Trusted by teams in health, financial services and the charity sector
Security, risk and compliance keep asking questions your team cannot answer with evidence. Every review adds weeks.
The EU AI Act, ISO/IEC 42001, your sector regulator and accessibility law. You know they matter. You do not know what they mean for this system.
Building on Azure or Databricks is one skill. Building something an auditor, a regulator and a board will accept is another.
Start here
A fixed-scope, fixed-fee review of one system, built on our Responsible by Design™ framework. You get a clear answer to one question: what stands between this system and a confident production sign-off?
What you walk away with
Then, if you want us
Our Assurance-to-Deployment (A2D™) method closes the gaps the assessment finds. We work inside your stack and your team, and we leave the capability behind.
Find the gaps and agree what good evidence looks like.
Governance, monitoring and human oversight designed into the architecture.
Senior engineers who write production code next to yours.
The documentation pack your approvers need, then go live.
| Criteria | Large consultancy | Development agency | FatFish |
|---|---|---|---|
| Understands AI regulation and assurance | Yes | Rarely | Yes |
| Writes production code on your platform | Rarely | Yes | Yes |
| Senior people doing the work | Partly | Partly | Always |
| Fixed-fee way in | Rarely | Sometimes | Yes |
| Leaves your team more capable | Sometimes | Rarely | By design |
Selected work
A research-intensive university had 450 live policies to test against the Higher Education (Freedom of Speech) Act before an Office for Students deadline. A human-in-the-loop review returned marked-up Word documents, every change cited, in six working days. A manual legal review had been quoted at sixteen weeks.
Two NHS mental health trusts needed their policy estates compared, not sampled. More than 500 clinical and corporate policies were mapped for duplication, variation and gaps against regulator-aligned benchmarks, so governance teams could prioritise the review. The trusts are not named.
SOPHIA already held policies and compliance evidence. We added an AI module on Azure that drafts standard operating procedures from source documents, asks when detail is missing, and leaves sign-off with a person. It was proved in a secure test portal, then built into SOPHIA.
FatFish Digital have been outstanding partners - their AI integration has transformed SOPHIA into a smarter, faster, tool for managing governance at scale.
No. We work inside your team and your stack. The aim is that your people can run and evidence the system after we leave.
AWS, Azure and Databricks, plus the usual surrounding tooling. If your stack is different, ask on the call.
We start from what applies to you: typically UK GDPR, the EU AI Act where relevant, ISO/IEC 42001, your sector regulator's guidance, security standards and accessibility requirements.
You own the roadmap. Use it with your own team, or bring us in to deliver it on a retainer. There is no obligation either way.
[Typical lead time, e.g. within two weeks of agreeing scope.]
The team
AI-driven technologist David founded FatFish Digital in 2013. He has since architected 100+ products including co-founding recommendations platform Upmost. He guides businesses through complexity to deliver standout digital solutions.
Marketplace and health-tech specialist, advisor and investor. Co-founded leading home improvement marketplace Rated People. Chairs UK therapist matching service Welldoing.org and workflow optimisation platform Carradale Futures.
Owais is a software architect and engineer focused on building scalable, real-world systems that solve meaningful problems. He has led complex full-stack and AI-driven projects at Fat Fish, specialising in LLM integrations, data pipelines, APIs, and cloud infrastructure. Known for his collaborative approach and calm problem-solving style, he brings clarity and positive energy to every team.
Philip is a product-driven strategist focused on solving user and client challenges. He leads cross-functional teams to deliver scalable, high-performing products throughout the full lifecycle. He guided Welldoing’s platform rebuild, managing requirements, timelines, and releases, and continues to optimise the platform based on user insights and analytics.
Thirty minutes with a senior consultant. Tell us about the system; we will tell you where we would look first.