In partnership with

This week’s theme is behaviour change. Firms are no longer just approving AI tools, they are actively pushing adoption and packaging it for clients. Shoosmiths hitting a 1,000,000 prompt milestone with Microsoft Copilot, tied to a firmwide bonus pot, is a useful case study in promotion of AI tools.

Additionally, an AI only employment service is getting headlines, and a tribunal judgment correction has reignited the accuracy debate.

AI in Practice

Shoosmiths has linked part of its firmwide bonus pot to Microsoft Copilot usage, with a target framed in prompt volume, and reports that it has now hit 1,000,000 prompts, triggering an additional £1,000,000 into bonuses. (Artificial Lawyer)

The immediate “so what” for UK firms is not whether Copilot can tidy emails or summarise notes. This scheme is what happens when you treat usage as the success metric. Prompt volume is easy to count, but it is a weak proxy for client value or risk reduction.

The more interesting question, in this author's view, is governance. If people are incentivised to use AI frequently, there must be guidance on what tasks it is suitable for and what information can be put into the tool. There must be also be guidance on what human supervision is necessary. Without this you risk optimising for activity rather than quality, and normalising casual AI use in a profession where confidentiality and privilege is paramount.

Takeaways

Gamification can push people towards low value or risky usage of tools. If you incentivise AI use in your firm, you must tighten your supervision and information handling rules at the same time.

On your radar

  • Peggie judgment corrected following AI speculation: A corrected version of the Peggie Employment Tribunal (Scotland) judgment has been published. In the last week there was widespread speculation about AI use by the judge as quotes did not appear in the cited source judgment. Why it matters for UK lawyers: the headline is less “did a judge use AI” (which is unclear) and more “verification and provenance is key”. (Scottish Legal) (Corrected Judgment), (Speculation on Reddit)

  • Grapple claims a £30,000 settlement as an AI only employment service: Legal Futures reports that Grapple Law, which is not a regulated law firm, negotiated a £30,000 settlement for a user and took 15% of compensation. Why it matters for UK lawyers: this is a real time test of the regulatory perimeter and consumer protection issues. This author is keen to see how this sits with both the Mazur judgment and client expectations of what they are buying when buying "legal services". It appears, in fact, that this product is more akin to a software sale, not provision of legal services and there must be a question of how close to the line can the marketing be when selling such services. (Legal Futures) (Artificial Lawyer)

  • TLT launches “TLT Intelligence”, packaging AI workspaces and delivery for clients: TLT has launched a client facing AI offering spanning “workspaces” and “collaboration”. Why it matters for UK lawyers: This is illustrative of a growing tend amongst larger firms to produce packaged AI products that clients can safely plug into. (TLT)

For Review

Practice Prompt

Try this template prompt in a tool your firm safely uses to produce a 3x3 triage of contract review. This should produce quick highlights of concerns on a given contract as a starting point for your advice.

Remember to insert your specific needs and context within the { } to give the best possible results.

You are a UK solicitor. Goal: {goal}.
Context: {facts}.
Constraints: {constraints}.
Deliverable: {deliverable} (tone: {tone}, length: {length}).
Verify: cite sources; flag uncertainty; list assumptions.

Task: Produce a 3×3 triage of {contract_type} for {client_type}.

Output:
1) 3 biggest risks with clause references and why they matter
2) 3 negotiable items with suggested compromises
3) 3 redlines with one-sentence replacement wording for each
4) 3 short questions to clarify commercial intent.

Standards:
- Keep to one screen of text if possible.
- Use clause numbers or headings for precise references.
- Mark assumptions explicitly.

How did we do?

Reply and tell me what you would like covered in future issues.

Serhan
Editor, UK Legal AI Brief

If you are finding this newsletter useful, consider subscribing to broader AI news. Superhuman AI, a newsletter I subscribe to, is advertised below.

Become the go-to AI expert in 30 days

AI keeps coming up at work, but you still don't get it?

That's exactly why 1M+ professionals working at Google, Meta, and OpenAI read Superhuman AI daily.

Here's what you get:

  • Daily AI news that matters for your career - Filtered from 1000s of sources so you know what affects your industry.

  • Step-by-step tutorials you can use immediately - Real prompts and workflows that solve actual business problems.

  • New AI tools tested and reviewed - We try everything to deliver tools that drive real results.

  • All in just 3 minutes a day

Disclaimer

Guidance and news only. Not legal advice. Test outputs and apply professional judgment.

Keep Reading

No posts found