The Foundation Before the Furniture – David Ricketts
AI readiness for a law firm means having clean, well-permissioned data, a pricing model that doesn’t lose revenue when work gets faster, and staff trained to use the tools already licensed. For mid-sized UK law firms, fixing these three foundations usually delivers more return than buying another legal AI product from a crowded market.
Key takeaways
- Fix data before buying tools. AI search exposes poor permissions and messy matter data. Microsoft’s own Copilot guidance puts removing oversharing first.
- Fix pricing before measuring ROI. If AI cuts the hours on a task and you still bill by the hour, revenue falls. Fixed fees now cover the majority of matters at UK and Irish firms (Clio, 2026).
- Train people on what you already own. Most mid-sized firms underuse Microsoft 365 and their practice management system (PMS). Training is cheaper than a new vendor.
- The SRA is watching. Its warning notice of 17 August 2026 makes firms and supervisors answerable for AI errors and confidentiality breaches.
- Mid-sized firms have an advantage. They can move faster than large firms and don’t need to bet on unproven start-ups.
What does AI readiness mean for a law firm?
AI readiness is the state in which a firm’s data, systems, pricing and people can turn AI output into reliable, billable, compliant work. It describes the firm, not the tool. The same Copilot licence will produce useful answers in a well-governed tenant and risky ones in a messy tenant.
In practice a ready firm has four things:
- Matter data that is current, deduplicated and correctly permissioned.
- A pricing model that captures the value of faster work.
- Staff who know how to prompt, verify and escalate.
- A written policy that maps all of this to SRA duties.
Where are UK law firms with AI in 2026?
Most UK firms already use AI, but governance has not kept up. Clio’s 2026 State of Legal Tech report found that only 39% of UK law firms have strong AI policies and oversight in place. AspiraCloud’s survey, The State of IT in UK Law Firms, found that 33% are using AI tools with no formal policy governing how AI is used, what data goes into it, or who is accountable when something goes wrong. cliolegalfutures
The regulator has noticed. According to the Law Society Gazette, the SRA received 42 reports of AI-related wrongdoing between July 2025 and July 2026. mondaq
The supplier market is crowded at the same time. The Financial Times recently described a legal tech market of more than 5,000 companies. [Link the FT article and confirm the figure before publishing.]
Why do law firm AI projects fail to pay back?
They fail for three reasons that have nothing to do with the AI itself: weak data foundations, hourly pricing, and low adoption. Each is covered below.
1. Your legacy infrastructure will dictate your AI ROI
AI tools can only be as good and as safe as the data and permissions they sit on. If you connect an AI search tool to your document management system (DMS), email or Teams, it will surface whatever users can technically open. That includes old matters, duplicate drafts and folders that were shared too widely years ago.
2. Don’t buy software to fix a pricing problem
AI efficiency cuts revenue under pure hourly billing. As an illustration, if a review billed at five hours takes two with AI, the fee for that task falls by 60%. The client benefits, but the firm’s margin doesn’t. The fix is commercial, not technical.
3. Adoption fails in the classroom, not the server room
Licences don’t create value until people use them well. Many firms already pay for AI and automation features in Microsoft 365 and their PMS that nobody has been trained to use.
How do you get your data and permissions ready for Copilot or legal AI?
Start with a permissions audit before anyone gets a licence. Then clean and label your content, and pilot with a small group. Microsoft 365 Copilot retrieves files, emails and chats using each user’s existing Microsoft 365 permissions, so any oversharing becomes instantly searchable. greenhood
This is a common reason for delay. A Gartner survey of 132 IT leaders in June 2025 found that data oversharing prompted 40% to delay their Copilot rollout by three months or more. comparethecloud
A practical sequence:
- Find overshared sites and libraries. SharePoint Advanced Management’s Data Access Governance reports show which sites are overshared and why. bdemerson
- Contain the risk during the pilot. Microsoft suggests using Restricted SharePoint Search during a Copilot pilot to narrow the sites Copilot can draw on while you review permissions. microsoft
- Label and protect client content. Microsoft Purview sensitivity labels classify content by sensitivity, and Purview DLP can block or audit sensitive prompts in Copilot. microsoft
- Check your DMS and PMS. Ethical walls, matter-level security and retention rules in your DMS and PMS need the same review as SharePoint. This matters most when you add AI connectors.
- Pilot, audit, expand. Start with one practice group, review what Copilot surfaced, and only then widen access.
Read more: Copilot for Legal: Why UK Law Firms Need More Than a Licence
How will AI affect hourly billing and fixed fees?
AI is accelerating the move away from the billable hour. Firms that move suitable work to fixed or value-based fees keep the margin that efficiency creates.
Clio’s UK & Ireland Legal Insights Report 2026 found that fixed or flat fees now account for 53% of matters across UK and Ireland firms, while hourly billing has dropped to 32%. A LexisNexis survey of 543 UK legal professionals found that 55% believe AI will change law firm billing models, up from 40% in January 2025. legalfutureslexisnexis
What to do:
- Identify the repeatable, high-volume work where AI saves the most time. Typical examples are conveyancing stages, first-draft contracts, disclosure triage and client onboarding.
- Price that work as fixed fees, using your own time-recording data to set the price.
- Keep hourly or capped fees for work that is genuinely unpredictable.
- Track margin per matter type, not hours per fee earner.
What AI training do solicitors and support staff need?
Training should be role-specific, practical and recorded, and it should cover verification, confidentiality and escalation, not just tool features. The SRA expects supervisors to make sure AI-assisted work is properly checked. It expanded its guidance on effective supervision in relation to AI in June 2026. solicitorsjournal
A sound training programme covers:
- Verification. Checking every citation and authority against the source.
- Confidentiality. Which tools are approved for client data, and which are not.
- Escalation. When AI output needs a second review.
- Hands-on use. Practical sessions on Copilot, Outlook, Teams and PMS automation that staff already have.
- Refreshers. Repeat training whenever a new tool or feature goes live.
Competence standards are also changing. In October 2026 the SRA said it wants to update its competence standards to reflect the growing use of technology and AI, with the revised statement due to be published in April and come into force in September. law360
What does the SRA expect from firms using AI?
The SRA expects the same professional standards whether or not AI is involved. Firms must be able to show the checks, training and supervision they had in place. The SRA’s warning notice was published on 17 August 2026 and applies to SRA-regulated firms and individuals. The SRA says it will take the notice into account when exercising its regulatory functions, and if something goes wrong a firm may need to explain what checks, training and supervision were in place. jmw
The notice highlights two risks, AI-generated inaccuracies and the exposure of confidential client information, and makes clear that those responsible for supervision may themselves face consequences. The SRA isn’t telling firms to stop using AI. It is telling them to govern it. vinciworks
Do UK law firms need an AI policy?
Yes, in practice. Without a written policy, a firm can’t show the SRA what checks, training and supervision were in place. The Law Society’s guide Generative AI – the essentials includes a checklist. It asks firms to define the purpose and use cases of the tool, outline the desired outcome, and follow their obligations under the SRA Code of Conduct, Standards and Regulations and Principles. lawsociety
A workable policy covers:
- the list of approved tools
- the data that may go into each tool
- verification rules
- who signs off AI-assisted work
- client disclosure
- training requirements
Should you buy a specialist legal AI tool or use what you already own?
For most mid-sized firms, the answer is to start with the AI already in Microsoft 365 and your PMS. Add a specialist platform only where a clear use case and budget justify it.
| Option | Best for | Main risk | Readiness needed first |
|---|---|---|---|
| Microsoft 365 Copilot | Email, Teams, drafting and search across documents the firm already holds | Surfaces overshared content | Permissions audit, sensitivity labels, a pilot group |
| AI features in your PMS or DMS | Matter workflows, time recording, document assembly | Features go unused | Training and workflow redesign |
| Specialist legal AI platform | High-volume review, research or drafting in one practice area | Cost, lock-in, vendor stability in a crowded market | A defined use case, a pricing model to capture the value, due diligence on data handling |
| Free consumer chatbots | Not suitable for client work | Confidentiality and data-retention breaches | Should be blocked or restricted by policy |
Read more: Copilot Without the Lock-in: How Law Firms Can Adopt Microsoft AI on Their Own Terms
What does AI readiness cost, and how do you measure ROI?
There is no reliable public benchmark for what readiness costs a UK law firm, so be wary of any single figure. The cost depends on:
- the state of your tenant and DMS
- how many users need licences
- how much training you deliver
The main cost drivers are the permissions clean-up, labelling, licences and training time.
Measure ROI in terms partners already use:
- time saved per matter type
- change in write-offs
- realisation rate
- fixed-fee margin
- active-user adoption rates
For context, in Wolters Kluwer’s 2026 Future Ready Lawyer survey, 62% of legal professionals said AI tools already save them between 6% and 20% of their weekly working time. That time only becomes ROI if your pricing captures it. lawfuel
What should a mid-sized firm do first?
Work in this order: govern your data, fix pricing on the work AI speeds up, train people on the tools you already have, and only then evaluate new vendors.
- Weeks 1–4: Run a permissions and oversharing audit, publish an AI policy and an approved-tools list, and block unapproved tools.
- Weeks 4–8: Label sensitive content, pilot Copilot with one team, and identify the three matter types with the most repeatable work.
- Weeks 8–12: Re-price those matter types, roll out role-specific training, and measure adoption and margin.
- After that: Evaluate specialist tools against a written use case and data-handling checklist.
The mid-market advantage
Mid-sized firms can adopt AI faster and more safely than larger competitors, because they can change their data, pricing and training without years of committee approval. You don’t need to bet on the next 5,000 start-ups. Getting the foundations right turns the tools you already pay for into a competitive advantage.
FAQ
What is AI readiness for a law firm?
AI readiness is the point at which a firm’s data, permissions, pricing and people can use AI safely and profitably. It covers clean, well-permissioned matter data, a pricing model that captures efficiency gains, trained staff and a written AI policy. It describes the firm, not any particular tool.
Is Microsoft 365 Copilot safe for law firms?
Copilot respects existing Microsoft 365 permissions, so it is as safe as your permissions are. If client folders are overshared, Copilot can surface them to the wrong people. Audit and fix sharing first, label sensitive content with Microsoft Purview, and pilot with a small group before a wider rollout.
Do UK law firms need an AI policy?
In practice, yes. The SRA’s August 2026 warning notice says firms may need to explain the checks, training and supervision they had in place if something goes wrong. A written policy covering approved tools, permitted data, verification and sign-off is the simplest way to show that.
What did the SRA’s AI warning notice say?
Published on 17 August 2026, the notice reminds SRA-regulated firms and individuals that using AI does not change their professional obligations. It focuses on AI-generated inaccuracies, such as false citations, and the exposure of confidential client information. It also warns that supervisors may face consequences for inadequate oversight.
Will AI end the billable hour?
Not overnight, but it is speeding up the shift to fixed fees. Clio’s 2026 UK & Ireland report found fixed or flat fees on 53% of matters, against 32% billed hourly. Firms that keep billing efficient AI-assisted work by the hour will see revenue fall on that work.
Should a mid-sized firm buy a specialist legal AI platform?
Only after the foundations are in place and there is a clear use case. Most mid-sized firms get faster returns from AI features they already license in Microsoft 365 and their practice management system. Specialist platforms make sense for high-volume, well-defined work where the pricing model captures the value.
What AI training do solicitors need?
Solicitors and support staff need practical, role-specific training on verifying AI output, protecting client confidentiality, and knowing when to escalate. It should be recorded and repeated whenever tools change. The SRA has added AI material to its supervision guidance and plans to update its competence standards.
BIO – https://www.linkedin.com/in/david-ricketts-7437991/
Sources:
- Clio UK: AI in UK Law Firms
- Legal Futures: The AI governance gap in law firms
- Gateley: The SRA’s AI warning notice
- JMW: SRA AI warning notice
- VinciWorks: Is SRA’s AI warning a wake-up call?
- Law360: SRA targets AI hallucination risks
- Solicitors Journal: SRA cautions on AI use
- Law Society: Generative AI – the essentials
- Legal Futures: Clio UK & Ireland Legal Insights Report 2026
- LexisNexis: Will AI change law firm billing models?
- LawFuel: UK law firms and the billable hour
- Microsoft Tech Community: Limiting Copilot data exposure
- Compare the Cloud: Stop Copilot surfacing confidential data
- BD Emerson: Microsoft 365 Copilot security
- Greenhood: Copilot permissions audit