The Forbes story this week confirms what we've been seeing across UK law firms, insurers, and accountancy practices: legal AI has reached a capital inflection point. £2.1 billion invested globally in the first half of 2026. Products like Soxton are now combining AI-generated drafts with licensed attorney review to lower costs on routine work. This is real. It is working. And it is creating genuine competitive pressure on mid-market practices that have not yet figured out their own AI strategy. For any regulated firm still operating without a clear position on AI-assisted work, this headline should land as a call to action, not a comfort.

But here is what the investment number obscures: capital flowing into legal AI does not equal capital flowing into good legal AI. The pattern we are seeing across UK professional services is a race to the middle—everyone adopting versions of the same AI products (Harvey, Legora, Luminance, Copilot for Microsoft 365) without the governance infrastructure to know whether the AI is actually working correctly. Many firms have bolted AI onto existing workflows without rethinking the workflow itself. They have bought the tools. They have not built the discipline. The SRA Code 2024 expects regulated firms to take 'reasonable steps' to ensure that their use of AI is competent and that material risks are identified and managed. Many firms are not doing either. They are treating AI adoption as a box-ticking exercise in response to competitive noise.

Trovix's perspective is direct: the problem with most legal AI implementations is not the model. It is the absence of rigorous governance. When Soxton combines AI generation with attorney review, that model works because there is a human quality gate. But most firms implementing AI at scale are not building equivalent gates. They are hoping the AI is right, not proving it. This is why Trovix Audit exists—not to replace human judgment, but to create transparent visibility into what the AI is doing, where it is confident, and where it is not. A firm deploying Harvey or Luminance without an audit trail, a confidence score, and a documented review process is not reducing risk; it is redistributing it. The investment story this week will accelerate adoption. The regulatory consequences will follow. The firms that survive will be those that invested in governance first and tool adoption second.

For any mid-market law firm, insurance broker, or financial services outfit reading this: you do not need to wait for the next funding round to compete. You need to make three decisions now. One: which routine, high-volume, low-discretion work will you automate first? (For most firms, this is client intake, document production, or regulatory screening.) Two: what does successful AI output look like for that work, and how will you prove it? (This is where most firms fail.) Three: what does your team need to know to work alongside AI without losing judgment? (Training is not optional.) If you cannot answer all three, do not deploy the AI yet. The cost of getting it wrong—in regulatory terms, in client outcomes, in reputational damage—is higher than the cost of waiting another quarter to get your governance right. Trovix Watch can help you track the regulatory signals as the SRA, FCA, and ICO clarify AI compliance expectations. But the real work is internal.

Source: Forbes