About FireBreak

AI risk testing, led by people who have built and run AI in production.

We know how AI fails in the real world because we've run it there. Now we use that experience to test other organisations' AI, independently.

FireBreak's leaders in conversation with a client executive in a boardroom

Independent by design

We never assess systems we built, and we never sell fixes for what we find.

So we have no reason to find problems that aren't there, and no reason to miss ones that are. We give clear guidance your own team or any provider can act on, and we can re-test once changes are made.

Proven in production, not in slideware

Live in national public services

Our AI is in production for NHS 24 and the Scottish Government, where accuracy, fairness and public accountability are tested every day.

Chosen over building in-house

Thoughtworks, a global technology consultancy, chose us rather than building in-house.

On the Scottish Government's AI Register

The first external company on the Scottish Government's AI Register.

A method you can verify

Our five-step method is repeatable, and the test harness lets anyone check the results independently.

“FireBreak told us plainly where our AI plans carried risk and what had to be in place before we scaled. That independent view gave our leadership team the confidence to move forward.”

Gary Sage, Chief Operating Officer, Crimson

Leadership

Jora Gill

Co-founder, FireBreak

There is a significant difference between AI that appears impressive in presentations and AI that can be confidently deployed.

Jora has been Chief Digital Officer at The Economist and SHL, and CTO at S&P and RELX. He sits on FTSE boards and has served on government boards. In these roles he has led and advised on AI, cyber security and risk.

As The Economist's first Chief Digital Officer, he led its data science team and drove the transition from print to digital, launching the Espresso daily briefing app. At SHL, he led AI initiatives across its assessment platforms. He has also held senior technology roles at Elsevier, S&P Global Ratings and BNP Paribas.

As a board member, Jora is accountable for technology decisions, investment and risk. He knows what a board needs to see before it can trust an AI system, and that is the evidence FireBreak's assessments are designed to provide.

Gurdeep Grewal

Co-founder and Chief Commercial Officer, FireBreak

Investing in technology is relatively straightforward. Ensuring it delivers tangible commercial outcomes is far more challenging.

Gurdeep has spent more than 40 years building and leading technology businesses internationally.

He was CEO of intive, a global design and engineering firm headquartered in Munich, where he tripled revenue and led three strategic acquisitions. As Senior Vice President and General Manager of GlobalLogic's European operations, he drove consistent double-digit growth before GlobalLogic was acquired by Hitachi. His career also includes senior roles at Zensar Technologies, CGI (Logica), HP and Unisys.

For Gurdeep, AI is only worth what an organisation can prove about it: where it creates value, where it carries risk, and what has been done about both.

Kay Ridge

Chief Marketing Officer, FireBreak

Kay has more than 20 years of experience in marketing, business development and strategic advisory. She helps FireBreak explain complex AI risk in clear terms that boards, regulators and clients can act on.

The specialists who do the testing

The testing is carried out by specialists who have built and run AI in production, alongside a CREST-accredited security partner. None of them built the system they're testing.

Security

AI red teamers

Attack your AI the way a real adversary would, through prompt injection, jailbreaks, data extraction and misuse of agent permissions.

Reliability

AI evaluation engineers

Measure accuracy, consistency and grounding against your real use cases.

Fairness and Inclusion

Fairness specialists

Test whether outcomes differ across groups, and why, using your own data where lawful or representative test sets where not.

Governance

Governance specialists

Review ownership, oversight, audit trails and incident response, and assemble the evidence pack.

Find out where your AI stands in two to four weeks.