Most UK firms have deployed agentic AI without governance frameworks in place—a regulatory blindspot that will become expensive. The question is not whether to govern AI, but whether you can prove you already are.
AI Governance  Trovix SiftLegal · Financial Services · Insurance · Accountancy

The Red Hat survey published in April 2026 cuts to the heart of a silent crisis in UK professional services: 87% of IT leaders deploy agentic AI systems, yet only 25% have strong governance in place. For law firms, insurers, accountants and financial advisers, this is not merely a technical problem. Under FCA Consumer Duty PS22/9, SRA Code Part C, and the PRA's expectations on model risk (SS1/23), governance of AI systems is not optional—it is a regulatory obligation. When less than half of surveyed firms have complete visibility of where their data is stored, processed and accessed, you are looking at firms operating in material breach of their own regulatory requirements, whether they know it or not.

This gap exposes a pattern the industry is only now confronting. Most UK firms have bolted AI onto existing workflows without first asking what data enters the system, where it goes, who can access it, and whether the tool can be audited. They have bought tools—Harvey for legal drafting, Luminance for contract review, Microsoft Copilot for general work, Legora for document handling—without building the governance layer that makes those tools safe in regulated environments. The tools themselves are capable. The problem is institutional: governance has been treated as a post-deployment concern, not a prerequisite. In a sector governed by the ICO, FCA, SRA and PRA, this order is backwards.

Trovix's approach is built on a different premise: governance and capability are inseparable. We do not deploy document intelligence, RAG assistants or client-facing AI without first establishing what data is in scope, who controls it, and how the output will be audited and retained. Trovix Audit is designed specifically to answer the questions that this Red Hat survey shows most firms still cannot: What is your AI doing? What data is it seeing? Can you prove it complies with your obligations? Other vendors focus on feature richness. We focus on defensibility. For a mid-market law firm or insurance broker, that distinction matters. A tool that is powerful but opaque will eventually create a compliance incident. A tool that is transparent and governed will not.

If you lead IT or compliance at a UK professional services firm, act on this now. Do not wait for your regulator to ask. Audit every AI system currently in use—including the ones deployed informally by departments. Map your data flows. Establish who owns the output and who is accountable if it is wrong. If you cannot answer those questions, take the system offline. Then, when you re-implement, build governance first. Trovix Sift and Trovix Aria are built with this model in mind: they are designed to work within documented, auditable data governance frameworks from day one. This is not a technology problem that technology alone solves. It is an institutional commitment that separates firms that will pass regulatory scrutiny from those that will not.

Source: Computer Weekly

Related Trovix product:

Trovix Sift →Book a demo →