Vertafore's new Velocity AI agents—covering policy servicing, document research and routine operations—represent exactly what should happen in financial services software: vendors building domain-specific automation into tools where they already hold the data relationship. For independent insurance agencies, this is practical. The story matters to UK mid-market firms because it shows the template now being copied across legal tech, accountancy software and financial services platforms. But here is what Vertafore's announcement does not say: who owns the compliance risk when these agents make a mistake? Who audits what the agent decided? That silence is the real story.

We are in the phase where AI vendors build the agents; compliance teams are still figuring out what 'responsible deployment' means in practice. Harvey, Legora and Luminance have all shipped document AI tools that work. But the market is now splitting into two camps: vendors who embed governance assumptions into their agents (limited autonomy, explainability baked in, audit trails mandatory), and vendors who ship autonomy first and assume the firm will bolt governance on afterwards. Vertafore sits in that second camp. The FCA's Consumer Duty PS22/9 and the approaching EU AI Act are making the first approach the only one that survives regulatory scrutiny. Firms using agents without that architecture are building liability.

Trovix's position on agentic AI is direct: agents are useful. But they are not useful if they operate without continuous visibility. This is why we built Trovix Audit as the necessary counterpart to any agent deployment—not as an add-on, but as the control system. The moment a firm says 'we are deploying AI agents', the next sentence should be 'how do we know what they did and whether it was compliant?' Vertafore's agents do real work. But a mid-market insurance broker, law firm or accountancy practice implementing them needs to answer: can we prove to the FCA, SRA or PRA that the agent's decisions were auditable, traceable and within risk appetite? Most firms cannot. That is a governance gap, not a tool gap.

If you are running a mid-market regulated firm and watching the agent wave arrive, do not wait for your software vendor to solve governance for you. Audit and explainability tooling should be selected now, before you deploy agents at scale. The cost of bolting it on later—or discovering during an FCA compliance visit that you cannot explain what your agents decided—is far higher than building it in. Vertafore's product is good. Your governance framework needs to be better.

Source: Fintech Global