Bloomberg's June story on AI transforming billing at top law firms describes a genuine inflection point: the billable hour, which has anchored legal business models since the 1950s, is finally breaking. Top firms are using AI to handle routine tasks faster, which means fewer billable hours to sell. This is not speculation about future impact—it is happening now at firms large enough to absorb the disruption. For mid-market UK legal practices, accountancy firms and financial services operations, this matters urgently. If your competitors are already deploying AI to compress time spent on standard work, you are not competing on efficiency anymore—you are competing on whether your billing model survives at all. The SRA Code of Conduct recognises that fee arrangements must be clear and fair; what the Regulator has not yet fully addressed is how practices should transparently communicate the impact of AI-driven productivity gains to clients. That gap is closing fast.
This story is part of a pattern that reveals the industry is in the early stages of a three-part shift. First, firms are using AI to kill cost on high-volume, repeatable work—document review, contract analysis, legal research, time recording. Second, they are discovering that AI's real business value is not cost reduction but quality improvement and risk reduction, which clients will actually pay for. Third, and most dangerously, they are learning that the practices which invested in AI earliest have a structural advantage: they understand their own data, their own workflows, and their own risks in a way that practices still using spreadsheets and legacy systems simply do not. The firms using Harvey, Legora or Luminance are not just faster—they are collecting operational intelligence that competitors cannot easily replicate. For mid-market firms without that infrastructure, the clock is ticking.
Our view at Trovix is direct: the problem with how most law firms, insurers and accountancy practices are approaching AI right now is that they are treating it as a cost-reduction tool first and a compliance and governance problem second. That is backwards. Products like Trovix Aria work because they start with what your firm actually needs to demonstrate to the FCA (Consumer Duty PS22/9), the SRA, or the ICO about how AI is being used, and what data is being processed. You cannot implement AI responsibly without that foundation. Many firms are deploying Copilot or general-purpose AI assistants without mapping them against UK GDPR, the EU AI Act's compliance requirements, or ISO 42001 governance frameworks. The result is unmanaged risk buried inside productivity gains. The practices that win in 2026 and beyond will not be the ones that move fastest on AI—they will be the ones that move smartest on AI governance. That means audit trails, permission matrices, and documented decision-making. Trovix Audit exists because mid-market firms need that visibility without hiring a separate AI compliance team.
If you run a mid-market legal, insurance, accountancy or financial services firm, here is what to do this quarter. First, map the work currently billed by the hour and understand which parts AI can genuinely compress—contract review, due diligence, document extraction, research. Second, do not assume your existing billing model will survive the transition; explore alternative fee structures with your clients now, before they demand it. Third, and most urgent, audit your current AI deployments. If you are already using ChatGPT, Copilot or any LLM with client data, you need to know what is being sent, where it is stored, and whether your terms of engagement allow it. Fourth, invest in one governance layer—whether through Trovix or another SRA and FCA-aligned approach—before you add more AI tools. The cheapest time to get governance right is before you have seventeen different AI systems running across practice.
Source: Bloomberg