The short answer
Lawtech covers technology that supports, supplements or replaces traditional methods of delivering legal services. In UK firms adoption has concentrated on automating internal processes rather than client-facing services, and the SRA’s research identifies small firms as least likely to adopt, with lack of capital the biggest barrier. Investment is growing: UK lawtech attracted a record £116.6 million in 2025 across 295 active businesses. The obligations that govern adoption are the existing ones, particularly confidentiality, competence and supervision, and they apply at procurement rather than after deployment.
Address conduct risk at procurement
The Law Society’s practice note is explicit that regulatory and professional conduct challenges should be addressed at the design or procurement stage through comprehensive risk assessment. A tool bought on a demonstration and assessed for compliance afterwards is the wrong sequence, and it is the common one.
What lawtech actually covers
The Law Society defines lawtech as technologies that aim to support, supplement or replace traditional methods for delivering legal services, or that improve the way the justice system operates.
That definition is broader than the AI conversation that dominates coverage. A case management system, an e-signature platform and an anti-money laundering verification tool are all lawtech, and for most firms they matter more day to day than anything generative.
The distinction that determines the regulatory analysis is between tools that support a solicitor’s work and tools that deliver something to a client without a solicitor reviewing it. The Law Society has noted that wholly automated legal services, with no review or supervision by a solicitor, raise conduct risks it considers unclear how to mitigate, and that such products have not come to market in the way once anticipated.
Where adoption has reached
The market is growing and the distribution is uneven, which matters if you are deciding whether your firm is behind.
UK lawtech investment reached a record £116.6 million in 2025, with the number of active lawtech businesses rising from 270 in 2024 to 295, as we reported in our coverage of record UK lawtech investment.
Within firms, the SRA’s research paints a more modest picture than the investment figures suggest. Adoption has focused mainly on creating efficiencies by automating legal processes rather than on client-facing services, though there are signs that is beginning to change, with firms planning more interactive websites and online portals for keeping clients updated.
Nearly half of SRA-regulated firms undertook delivery innovation in the preceding year, and nearly half of that required new technology. The regulator also identifies the structural constraint plainly: small firms are the least likely to adopt lawtech, and the biggest barrier to adoption at scale is lack of financial capital, with investment tending to flow to larger firms serving corporate clients.
The categories worth understanding
Rather than a product list, the useful frame is what each category does and what it puts at risk.
| Category | What it does | Principal risk |
|---|---|---|
| Practice and case management | Matter tracking, time recording, workflow | Data migration and lock-in; it holds everything |
| Document automation | Generates documents from templates and data | Template error replicates across every matter |
| E-signature | Electronic execution | Execution formalities for particular documents |
| Legal research | Access to case law and legislation | Currency, and reliance without verification |
| Onboarding and AML verification | Identity and source of funds checks | The firm retains the regulation 18 obligations regardless |
| AI drafting and review | First drafts, extraction, review at volume | Accuracy and confidentiality |
| E-discovery and analytics | Large-scale document review | Cost, and whether review capacity matches the case |
Two of those risks deserve emphasis. A document automation template containing an error does not produce one bad document; it produces every subsequent document containing that error, which makes template governance more important than the tool. And an AML verification tool performs checks, but the firm remains responsible for the risk assessment and the underlying obligations, as our guide to what regulation 18 requires sets out.
How to choose without being sold to
Start with the problem rather than the product. Identify where work actually breaks down: manual document handling, onboarding delays, missed deadlines, time recorded late or not at all. A tool bought without a defined problem tends to become a subscription nobody owns.
The questions worth putting to any vendor handling client material:
- Where is data processed and stored, and under which jurisdiction
- Is client input used to train any underlying model, and can that be excluded contractually
- What is the retention period, and can it be set to zero
- Who at the provider can access the data, and in what circumstances
- What happens to the firm’s data if the contract ends or the provider is acquired
- What does exporting everything look like in practice, and has anyone done it
- Does the arrangement satisfy UK GDPR obligations as well as professional confidentiality duties
Named products are easy to find and their relative merits change with each release. What does not change is the evaluation framework, and a firm that can answer those seven questions has done the work that matters. Be cautious of recommendation lists that do not disclose whether a commercial relationship exists between the publisher and the products named.
The cost nobody budgets for
Implementation, data migration, training and the productivity dip while people learn the system routinely exceed the licence fee in year one. A procurement decision made on subscription cost alone is comparing the smallest number in the calculation.
The obligations that apply
There is no separate lawtech rulebook. The duties that govern adoption are the ones that govern everything else, and four do most of the work.
Confidentiality. Paragraph 6.3 of the Code of Conduct for Solicitors requires client affairs to be kept confidential. Any tool receiving client material engages that duty, and the contractual position with the provider is what discharges it or fails to.
Competence. Providing a competent service now includes understanding the limitations of the tools a firm relies on. A fee earner who cannot describe how a system fails does not fully understand the work it is doing.
Supervision. Where work is produced by a system rather than a person, the supervision obligation does not fall away. The SRA’s framing is that firms should oversee technology much as a solicitor supervises a junior employee: the work can be delegated, the accountability cannot.
Systems and controls. The Code of Conduct for Firms requires effective systems, records demonstrating compliance, and identification of material risks. An approved tools list and a record of who authorised what are part of that.
Data protection sits alongside. The February 2026 amendments strengthened the ICO’s powers, including the ability to require production of a specific data protection impact assessment, which our guide to UK data protection after the DUA Act covers.
AI, and where the line sits
AI is useful for first drafts, extraction and review at volume. It is not a substitute for legal advice, and the courts have made the consequences of treating it as one explicit.
Two risks are distinct and both are live. Accuracy: fabricated citations have produced wasted costs orders and regulatory referrals, and the responsibility sits with the lawyer who filed the document regardless of what produced it. Our guide to fake AI citations in court sets out the case law.
Confidentiality: in R (Munir) v Secretary of State for the Home Department [2026] UKUT 81, the Upper Tribunal observed that putting client material into an open AI tool can amount to placing that information in the public domain. That makes an approved tools list a confidentiality control rather than an IT preference.
Firms need a stated position on which systems may receive client material, communicated before staff are working to a deadline. The wider picture is in our guide to AI in legal practice.
What it means for those qualifying
Technology fluency is becoming part of what firms expect, but the useful version is narrower than general enthusiasm.
What distinguishes a candidate is not knowing which products exist. It is being able to verify what a tool produces, recognise where it fails, and explain why a particular document should not go into a particular system. Those are judgement skills expressed through technology rather than technology skills.
There is also a training consequence worth understanding. Document review at volume was never valuable because the documents needed reviewing; it was valuable because volume builds the pattern recognition that lets someone spot an abnormal contract quickly. Where automation removes the volume, that recognition has to be built another way, which is a live question for qualifying work experience, as our guide to QWE discusses.
Before signing anything
- Define the problem the tool solves, in one sentence
- Run the seven vendor questions on data, training, retention and exit
- Budget implementation, migration and training alongside the licence
- Decide who owns the tool internally and who approves access
- Confirm whether client material will enter it, and whether that is permitted
- Record the decision and the risk assessment, since the systems and controls duty requires evidence
- Add it to an approved tools list that staff can actually find
Frequently asked questions
What is lawtech?
The Law Society defines it as technologies that aim to support, supplement or replace traditional methods for delivering legal services, or that improve the way the justice system operates. It covers case management, document automation, e-signature, research, AML verification, AI drafting and e-discovery.
How widely have UK firms adopted lawtech?
Adoption has concentrated on automating internal processes rather than client-facing services. Nearly half of SRA-regulated firms undertook delivery innovation in the preceding year, though the SRA identifies small firms as least likely to adopt, with lack of financial capital the biggest barrier.
What rules govern lawtech use by solicitors?
The existing ones. Confidentiality under paragraph 6.3 of the Code of Conduct for Solicitors, the duty to provide a competent service, supervision obligations where work is delegated to a system, and the systems and controls requirements in the Code of Conduct for Firms, alongside UK GDPR.
Can I put client documents into an AI tool?
It depends entirely on the tool’s contractual terms. The Upper Tribunal observed in R (Munir) v Secretary of State for the Home Department [2026] UKUT 81 that putting client material into an open AI tool can amount to placing it in the public domain. Firms need a stated position on which systems may receive client material.
When should compliance be considered in a lawtech purchase?
At the design or procurement stage, through comprehensive risk assessment, per the Law Society’s practice note. Assessing compliance after a tool has been bought and deployed is the wrong sequence and the common one.
Does an AML verification tool discharge the firm’s obligations?
No. The tool performs checks, but the firm retains responsibility for its firm-wide risk assessment, client and matter risk assessments and the underlying obligations under the Money Laundering Regulations 2017.
The key points
- Adoption is uneven: capital, not culture, is the main barrier for small firms
- Compliance belongs at procurement: not after deployment
- The duties are the existing ones: confidentiality, competence, supervision, systems and controls
- Template error scales: automation replicates mistakes across every matter
- An approved tools list is a confidentiality control: not an IT preference
The practical takeaway
The useful question is not which products to buy but what breaks down in the firm and whether a tool addresses it. Most disappointing implementations start with a demonstration rather than a problem.
Whatever is bought, two things decide whether it survives contact with regulation: whether the contractual position on client data is understood before signing, and whether someone owns the tool afterwards. Both are settled at procurement, and neither is expensive to get right at that point.