Everything is going agentic — AI that acts, not advises. Your product is a human-in-the-loop advisory layer. Are you strategically positioned for that wave or about to be flattened by it?

The full executive answer

My contention is that the agentic wave is the single best thing that can happen to this company, because agents industrialise exactly the risk we govern. An advisory AI that hallucinates costs you a bad slide; an agentic AI that acts on a hallucination costs you money, contracts or compliance breaches at machine speed. Every serious agentic framework — the EU AI Act’s high-risk obligations, NIST’s AI RMF, emerging agent-governance guidance — converges on the same requirements: human oversight at material decision points, logged rationale, bounded autonomy, and independent checking. That is not a description of our roadmap; it is a description of our shipped architecture: approvals engine for human sign-off gates, guardrail hard-blocks that stop non-compliant output from rendering at all, cross-provider consensus as independent checking, hash-chained audit as logged rationale.

So the strategic sequencing is: today, we govern AI-assisted human decisions; as customers deploy agents, the same control plane extends to govern agent-proposed actions — the approvals engine is precisely the "human authorises the consequential step" checkpoint that agent deployments will be required to have, and the platform’s asynchronous job and webhook infrastructure is the technical substrate an agent integration needs anyway. Current state versus roadmap, explicitly: the governance primitives are shipped and CI-tested; packaged "agent governance" as a marketed capability is roadmap, sequenced behind launch and first customers, and I will not pretend we have production agent deployments today.

The scenario where we get flattened is a world where organisations let agents make board-material decisions without oversight, evidence or approval gates. I am willing to bet the company that boards, regulators and courts will not permit that world — and if I am wrong about that, considerably more than this company is mispositioned.