Innovation Built on Trust
Why a trusted, governed foundation is not the brake on AI ambition in insurance and pensions. It is what makes bold innovation possible.
The foundation that makes AI work across insurance and pensions
If you are a leader in insurance or pensions, you are being asked to show a return on AI. This report is the case to bring to that conversation.
It starts where the boardroom conversation now sits: not what AI can do, but where the return is. It sets out why most pilots stall, and why the reason is rarely the model. It explains what makes insurance and pensions the hardest place to get AI right, and why that is an advantage for the firms that clear the bar. And it makes the argument at the centre of the whole piece: in a regulated business, governance is not the brake on AI. It is what lets you move fast without putting the business on the line.
88%
of organisations use AI in at least one function. Only 39% see any effect on operating profit. (McKinsey, 2025)
95%
of enterprise generative AI pilots deliver no measurable impact. (MIT NANDA, 2025)
65%
of European insurers already use generative AI, but most are still stuck at proof of concept. (EIOPA, 2025)
Six principle for AI in a regulated business
Governed data
Trusted data with known lineage and access, not uncontrolled copies, extracts and spreadsheets.
Auditability
Reconstruct what happened: which model, which data, which workflow, what output, what next.
Controlled authority
No agent has undefined power. Purpose, owner, permissions and escalation are clear.
Meaningful human oversight
Oversight placed where human judgement genuinely matters, not a tick in a box.
Model independence
Change the model without rebuilding the process around it. Models change; the architecture should not have to.
Runtime safety
Continuous monitoring and a clear way to stop an agent that acts outside its mandate.
Innovation Built on Trust
Why the foundation and not the model, determines whether AI delivers value in insurance and pensions
The people behind the thinking
John Reynders, CEO
The stakes, the strategic moves, and the view from the top.
Nelson Petracek, CTO
How we got here, and where it goes.
Iris Reynisdottir, Head of Product
Keylane Automate in practice, and what it opens up.
Nicolas Marzin, Head of AI
What governed AI concretely is, and where value lands.
Governance did not kill the car. It turned it into something billions of people trust their lives to every day. It is not a constraint on ambition. It is what makes ambition safe enough to scale. Iris Reynissdottir, Head of Product, Keylane
Put AI to work where it matters most
A trusted foundation you can prove. The freedom to build on it.