Introduction

I am a partner here at Ashford's and also a technology and data lawyer. So clients would pay me to advise on issues around technology licensing, data compliance, regulatory change, cybersecurity, and increasingly AI itself.

AI in the Legal Sector: Law Firms vs. In‑House

I'm speaking to you today, though, in relation to how AI is being used within the legal sector. And even within that, I think there's a distinction between how AI is being used by law firms as opposed to in-house teams in regular businesses.

Why Law Is Ripe for AI—And Why Adoption Lags

So AI and law are ready for change. The short answer is it depends on who you speak to. I think that there are a number of reasons that the legal sector is ripe for use of AI but culturally very resistant and hesitant to take up on it.

A Conservative Profession and Risk Aversion

We're a very slow-moving small C conservative profession. We're trained at law school in risk aversion. Clients pay us things about the worst case scenario.

Obviously with new technologies that can happen quite frequently And law firms have been around long enough to be skeptical around hype, bubble, trends. I've been doing technology law for 15 years, so I've seen crypto, blockchain, smart contracts, metaverse come and go.

From Hype Cycles to Staying Power

AI very much over the last couple of years, it's very clear it's here to stay. And now is the time. So thankfully, finally, we're backing the right horse here.

Work Characteristics That Suit AI

So despite all those opportunities for change, The work being repetitive, expensive, knowledge intensive and precedent based, there are still issues around uptake of AI.

Document-Heavy Matters: Litigation and DSARs

But having said that, the benefits or reasons for change, as I say, it's very document intense, lots of pages and documents being crunched in litigation or respect of data subject access requests when individuals have everyone has a right to know which information an organization holds about them typically we see that alongside employment grievances an employee is Leaving a company on not great terms, they're maybe doing a separate employment claim in the High Court or in the courts, and at the same time want to do a phishing expedition to know exactly the information their employer has on them, what's been being said.

Email and technology has been helpful in two ways here, in that it obviously enables us, gives us the solutions to crunch large amounts of data, review lots of documents, increasing with AI in terms of summarising or redacting those documents and getting them ready for court, but also in the sense that there's been a massive proliferation in information itself being created and held on people in the last 10 or 15 years.

Contract Negotiation Patterns and Playbooks

As well as the data trawling or disclosure, we're seeing contracts being marked up, negotiated, typically around the same points. So I know there are tech lawyers in the room. It will be things like indemnity, termination, liability, cybersecurity issues, IPR infringement coming up frequently.

Over a period of time, you get enough information from customers and feedback from documents to speak to a playbook or contractual negotiation guide around that. AI is very good at helping develop those, as I'll come on to.

Client Pressures: Fees and Diverging Expectations

One of the key things is client pressures on fees as well. Obviously, clients want to pay less. We're seeing a real divergence in the market, though.

Some clients are saying, we want you to use AI, but pass the cost benefit on to us, make sure you don't get it wrong. And the other clients are saying, we don't want you to use AI at all, we're very risk averse. So there's this kind of broad church that us as private practice lawyers continue to have to be able to service.

Talent, Retention, and Tech-Forward Culture

I think salary pressure as well is also quite relevant. We're using AI and legal tech as a recruitment and retention tool, particularly to younger lawyers, people coming out of university and law school who are much more technophilic, shall we say, than partners at the other end of the scale. For them, use of legal tech products is a real benefit and attractive factor in choosing law firms they apply to for their training.

And again, because we're working quite regularly on precedents, templates and structure, the data exists already to be crunched by AI. It's about making sure we have the right documents in place and the LLMs or issues to process those.

Barriers to Adoption in Private Practice

But as I say, there are a number of issues or factors to be aware of as well if you're looking to develop in legal or sell to legal firms.

Regulation, Insurance, and Confidentiality

It's very highly regulated. We're regulated by the Solicitor Regulation Authority and the Law Society. That basically says we have to keep client information confidential. The compliance standards are very serious.

The biggest expense any law firm is going to have is its annual insurance premium. And that varies taking into account how much law firm is using LegalTech as part of what they're providing to clients.

Data Security and Cyber Risk

Also, things around data security concerns. Obviously, law firms have lots of very juicy, highly sensitive information about their clients, be they individuals or enterprises.

As I say, I advise on cybersecurity. A lot of law firms are being targeted by ransomware and hackers at the moment.

Accuracy and Hallucinations

Hallucinations as well.

If AI doesn't work, the output isn't accurate. Who gets to blame? The lawyers, not the AI.

So that's something also that we have to think about and deal with.

The Economics Problem: Time-Based Billing vs. Efficiency

1The key reason for me around resistance, hesitancy around AI uptake in the private practice part of the legal sector is law from economics. All lawyers are trained to bill in six minutes units of time.

If I'm being presented with a solution that enables me to do a job more efficiently, law firms haven't yet worked out how to pass that value onto the client and also monetize it effectively. To my mind, if I'm using an AI tool that enables me to review a contract and accord for the time that would take me or a trainee to do it, I still, as a private practice lawyer, want to make sure I'm getting paid appropriately for the advice and expertise I'm using in respect of the AI overlay, the supervision I'm giving to that model. that doesn't always track through.

So I think that the efficiencies that AI provides in terms of time savings, law firms haven't yet worked out how to square that circle and monetize those benefits.

Why In‑House Looks Different

I think in house teams though, that's a different proposition.

Product Fit and Workflow Friction

client appetite for risk I mentioned, but one of the real reasons I think is that generally a lot of tools in the sector are still a little bit too clunky and inaccurate. They add an extra layer of bureaucracy or additional steps in the workflows that lawyers have to think about and they still require babysitting or oversight in many circumstances.

Cultural Hesitancy and Behavioural Change

One of the other key obstacles to change I think is cultural hesitancy and behavioural change. Chris mentioned this, I think that a lot of lawyers are used to working a particular way and I think it's very different now

or be very different in five years time training as a lawyer to how it was when i did it i guess 15 years ago let alone how people at the end profession or coming to the end of the profession did it kind of 30 years ago i think that in five years time junior lawyers will be as much legal engineers and users and evangelists for ai and legal tech as they are doing legal skills i think the effects of change in legal sector is going to be that

Evolving Roles: From Documents to Relationships

The job law becomes more about relationships with people as opposed to using the documents, the information to provide drafts or update step papers on checklists or provide regulatory advice or do research. It'll be dealing with people, delivering that message, dealing with the nuance and managing client relationships.

Current Uses of AI in Legal Practice

So what are we actually doing at the moment?

There's a lot of buzz in law around AI. I was at a fantastic legal tech conference last week.

I was saying to Chris earlier, it's basically billed as a Glastonbury of law fairs. I'm not sure how accurate that is, but there was karaoke and silent disco, so it went quite a bit of way towards that.

Document Review, Triage, and Summarisation

We're using AI across various firms or various firms using AI for things like document triage and summarising. If I'm doing due diligence, so my client is looking to acquire another company or invest in another company, they want to look over the hundreds of customer contracts that target company has to make sure there's nothing nasty lurking in there, there's no regulatory issues or IP infringement issues they're going to inherit when they buy that business. It makes sense rather than have an army of 40 lawyers spend two weeks look over the thousands of documents in that data room to run it through an LLM, say reviews documents, give me the key issues in each one, which ones are different, why they're different, what's the contractual term, what runs out where, what do we need to flag to the clients.

My feeling is that an effective AI product can do that with 90-95% accuracy of that team of 40 lawyers. So there are definitely cost saving benefits there. But again, the question remains how the law firms charge and monetize that.

Drafting, Clause Extraction, and Notes

We're also using AI for contract drafting, clause extraction, as most people are for call notes, annotations.

Firm‑Native Tools and Knowledge Bases

One of the real issues that law firms are facing from a regulatory sense is training LLMs on client data. So one of the ways around that is that you'll see firms have a native proprietary AI product, typically based on GPT, which trawls their internal document management system for knowledge about particular issues or clients and use that to create an internal data bank may be a chatbot style function.

Give me some insight into this particular agreement I worked on a few years ago. How does this client like their advice delivered? That kind of thing.

Onboarding and Compliance Automation

There's a bit of use of AI as well around the regulatory compliance side of things, when we're bringing new clients in, carrying out money laundering checks, that kind of stuff.

Marketing and Admin: The Low‑Hanging Fruit

I'm seeing a lot of AI generated marketing from law firms. It's almost as if there's that cartoon where AI is generating the content and reading the content at the other end and messages being lost, I think, in translation.

for me the real obvious use of ai in private practice is going to be the low hanging fruits so time recording is the bane of every private practice always life it's very relevant but we all hate doing it basics like email filing firm admin meeting recording etc those kind of things be really helpful if anyone is looking to build any products i'm very happy to assist with that

The Vendor Landscape

So these are some of the big players in legal tech at the moment. Top left one is CoCouncil, which is a Thomson Reuters product. Thomson Reuters are a big information management business in the legal sector. That's their flagship product.

I've got a demo I might run after this slide if I have time.

High‑Profile Platforms and Valuations

HeartV are making a lot of noise. People might have heard of this. They've recently been valued, I think, $5 billion.

I've trialled this product. It's very, very US-centric. Effectively, it's chat GPT, a bit of Anthropic, a bit of Claude in the back end, and aimed to be a kind of one-size-fits-all legal stop shop.

research me on this point, what is this firm doing, what is this client like, release documents for me, draft me an NDA. It does everything, and the idea is that I guess it's a pilot or desktop assistant. So as I say, $5 billion, I think that's a bit ridiculous, but we'll see how that develops.

Assistants and CLM Tools

Logora, by comparison, is the poor relation at just, I think, $675 million valuation. That's a Swedish business that started two years ago. So, again, provides a kind of AI assistant dashboard.

Juro is more contract collaboration and management.

These are the big players, as I say.

What Successful Adoption Requires

Remove Bureaucracy, Meet Lawyers Where They Work

But for me, I think the reason that AI uptake has been a bit mixed in the legal sector is the behavioural workflow changes. 1It needs to remove bureaucracy rather than add to it. It also needs to be seamless with those and make people's lives easier and come to us rather than us come to it.

Integrate with Email and Word

The tools that lawyers use 95% of the time are our email inbox and Word.

Selling into Law Firms: Stakeholders and Champions

Legal firms are also a very hard sell for AI and legal tech businesses. I think there's a risk of getting caught in a bit of a quagmire between an innovation team, a risk team, an infosec team, a marketing team, budgetary requirements. It may be the case that lawyers are very keen to use a particular tool, but rolling it out firm-wide is quite challenging.

And I think there's a particular route in, which is basically going to be having an internal champion who's sufficiently senior and empowered enough to convince the rest of the firm and relevant stakeholders that your AI solution is the best way forward.

Defining Success and ROI

So what would success look like?

Human Oversight, Time Savings, and Risk Reduction

Something that presents a seamless omnipresent tool, but still allows for human subdivision to enable us to deal with the risk issue and the client advisory piece, frees up lawyers time, doesn't add to the requirements, and the return on investment framed in hard numbers.

What's this going to save me in terms of time? How am I going to charge it to clients? How's it going to make me money? How's it going to avoid risk?

What’s Next: Emerging Models

So what's next? We're starting to see a few AI native law firms.

AI‑Native Law Firms and Public Tests

Garfield AI were featured on Channel 4 program earlier this week. There was a test between an AI lawyer and a real lawyer. I don't know if anyone saw that.

Basically, Garfield exists to enable people to do small value claims up to £10,000 in the High Court. It'll ask you what the situation is, what the fact pattern is to a particular issue, draft you the claim form, talk you through the process.

Training on Client Data and Regulatory Guardrails

LLMs trained and given free rein on client data as well. I think that needs to be linked up to regulatory oversight and guidance from the SRA and more AI native law firms, firms that are built around use of AI and technology as opposed to firms that use it as a bolt-on.

Conclusion

In summary, I think that selling AI to private practice firms is a bit of a tricky proposition, but I think conversely selling to in-house legal teams who want to be more efficient, free up legal budget, work more effectively is probably the way to go.

And I'd like to assist with that if anyone wants any help. I'm very happy to take questions.