The Law.com survey of 121 legal leaders tells a story that should worry any UK law firm still using AI as a margin protection device. In-house legal departments are now adopting AI to cut costs, insource complex work and reshape how they buy legal services. This isn't speculation or pilot projects anymore — it's operational reality for general counsels managing the FCA Consumer Duty PS22/9 compliance burden, navigating PRA SS1/23 governance, or handling the new EU AI Act implications. For mid-market UK law firms, this is the moment the pricing model breaks. When your client's in-house team can use systems like Harvey or Luminance to handle document review, contract analysis and due diligence at a fraction of what they paid you last year, the old 'we'll do it faster' argument collapses.

This is part of a deeper shift in how regulated firms think about technology investment. We're seeing the same pattern across insurance, financial services and accountancy: the firms that built AI strategies around cost-saving and risk reduction are now facing clients who did the same thing. The SRA Code of Conduct and ICO UK GDPR compliance frameworks mean that implementing AI poorly — without proper transparency, accountability or quality assurance — now carries real regulatory risk for both law firms and their clients. The firms winning right now aren't the ones claiming AI will solve everything. They're the ones honest about what AI can actually do: accelerate routine work, extract value from unstructured data, and free fee-earners to do work that requires judgment. The firms losing are still treating AI as a cost-cutting tool or a billable hour multiplier.

Where we differ from the current market approach is straightforward. Tools like Copilot and general-purpose RAG systems (including some of the high-profile legal AI products) work well for specific, narrow tasks but often fail in real practice because they don't understand the regulatory context of the work or the quality standards your firm's clients demand. Trovix Aria exists precisely because law firms, insurers and accountants need AI that's been trained on regulated firm workflows — not just generic legal documents. When a general counsel is deciding whether to insource work or keep paying external counsel, they're comparing the quality, speed and compliance confidence they get from in-house AI against what they get from you. If your AI is making errors or creating liability gaps, you've already lost that client. If your AI is genuinely better and genuinely defensible under the SRA Code and ISO 42001 frameworks, you have a reason to stay in the relationship.

Here's what a mid-market law firm, insurer or accountancy practice should do right now: First, stop thinking about AI as a defensive cost tool. Second, audit what your clients are actually using in-house — find out which work they're insourcing and why. Third, work with an AI implementation partner who understands regulated firm liability and audit trails, not one who's promising you generic productivity gains. Trovix Sift and Trovix Brief exist because we've seen hundreds of firms fail at AI because they picked tools designed for unregulated environments. Fourth, be transparent with clients about what AI you're using, how it's been tested and what quality standards you're holding it to. That transparency is now a competitive advantage because most firms still aren't doing it. The firms that survive the next 18 months will be the ones competing on actual quality and compliance confidence, not on who can promise the fastest cost reduction.

Source: Law.com