skip to content

AI in Action: Learnings from Legal Teams in the UK

by Vivan Marwaha

Artificial Intelligence is not new to the legal profession. Legal teams have been using the technology in ediscovery, due diligence, and contract review for a while. But generative AI has taken the world by storm in the last six months. As the legal industry begins thinking about this latest development in AI, Everlaw, in collaboration with UK’s The Lawyer, recently hosted a webinar on ‘AI in Action: Real-Life Examples of How Law Firms and Individuals Are Using AI Today.’

AI in Action webinar panelists

Panelists included Stacey Rickford, Legal Operations Director at Travers Smith, Danielle Costin, Projects and Innovation Director at Shoosmiths, Peter Lee, Partner and CEO at Simmons and Simmons, Matt Byrne, Director of Insight at The Lawyer, and Tom McKechnie Ward, Customer Success Manager at Everlaw. In this record-breaking webinar with nearly a thousand live attendees, the panelists engaged in a lively discussion on the topic on nearly everyone's minds these days: the impact of generative AI on legal teams. 

Early Use Cases – Summarization and Reasoning

Given the diversity of organizations the panelists represented – from leading law firms in England to legal service providers and technology platforms – the speakers had very different experiences with adopting generative AI. Peter Lee highlighted how Simmons and Simmons has started a new data science team, appointed a global leader for LLMs, and has been empowered by its board of directors to look at the several use cases for generative AI in their work.

The early conclusion? Summarization.

Generative AI is emerging as a powerful tool to summarize documents, particularly in classifying regulations. Although Peter mentioned that it has room to grow, the technology is particularly powerful when users enter the right prompts and have configured their questions correctly. 

Everlaw’s Tom McKechnie Ward, agreed, sharing his interactions with clients by highlighting the incredible value of AI in generating summaries for matters in litigation.

As data volumes continue to grow while resources remain flat, legal teams need to learn to do more with less, and generative AI can help save time and money by summarizing 500-page documents which may otherwise take countless hours of billable or unbilled time. The technology can be further utilized through prompts. Instead of running keyword searches on large databases, users can just prompt into the database instead, with prompts such as “these are the issues I’m investigating, can you suggest the thousand most relevant documents to my search,” Tom highlighted.

Travers Smith’s Stacey Rickford shared her firm’s early adoption of generative AI and took the discussion one step further by breaking down AI into two broad use cases: generative, where AI is generating and creating brand new content, and a second, which is in reasoning.

Reasoning is where Travers Smith is currently focusing its innovation and investment.

Reasoning has existed in some form of AI for years, where, as Stacey explained, “users train AI to recognize a particular piece of data through a series of data leveling.”

However, the difference today is that AI has become much more sophisticated as it has the benefit of everything on the internet. “So instead of a relatively small data set where you've trained it, now its reasoning is much, much broader,” Stacey continued. “And it does appear to have a high degree of accuracy.”

Currently, her firm is working on a case where a client is interested in investigating a breach of contract. The client wants to know whether there’s any evidence of intention to breach before it happened. The answer? AI. How? As Stacey Rickford explains, “what we're doing is looking at about 18,000 emails, and we're asking the AI to run through it and basically look for a smoking gun, look for any evidence of breach. And that's something that previously might have taken a trainee or a paralegal months. And potentially, the AI can do it very quickly."

AI is helping her team accomplish tasks that would have previously taken months in mere weeks or days.

Securing Client Data

No new technology comes without risk, and one of the biggest risks raised by the panelists was the security and privacy of client data. Given large language models are trained on vast amounts of publicly available data, the risk that sensitive client information being used and produced in a completely different setting is high. 

However, this risk is largely concentrated among users of the consumer, or free, versions of generative AI applications. Consumer-grade licenses and subscriptions do not contain any clauses for generative AI platforms to delete the data fed into their systems, and therefore pose significant privacy and confidentiality concerns. However, enterprise-grade licenses with these companies allow users to ensure their data, and by extension, client data is deleted immediately after usage. 

“Like many law firms, the very first thing we did was to block ChatGPT to make sure that some of our junior lawyers, in particular, weren't just copying and pasting emails that they'd received straight into it and hoping for a very quick answer. So what we did was we became B2B customers of OpenAI” explained Stacey Rickford. “When you have an enterprise license, and so you're a business-to-business customer, there is a contract in place between you and that software company where they will say that they will delete all of the data that goes into that,” she further detailed.

Since the enterprise version does not have a user-friendly interface, Travers Smith built their own interface and then open-sourced it for other law firms, legal teams, and other organizations to use. Called YCNBot (Your Company Name Bot), the code has been adopted by more than 200 companies, demonstrating considerable interest in the field, and a desire for legal teams to let their employees learn and adopt this technology in a safe manner. It is also important to note that some legal technology platforms have enterprise licenses with OpenAI and other generative AI platforms which ensure that their user data is similarly deleted. Read about Everlaw’s generative AI principles here.

An Opportunity for Transformation

Shoosmiths’ Danielle Costin reminded the panel and the audience that while it’s important to be aware of the threats we should acknowledge that this is a fantastic opportunity for transformation and the legal profession shouldn’t run away from it. 

The time savings from generative AI are going to be enormous, allowing in-house teams to achieve time-consuming tasks in incredibly short periods of time.

New types of legal services, such as legal prompt engineering may emerge, and new professions may get created to take advantage of generative AI. Thirty years ago, web designers didn’t exist as a profession because the web didn’t really exist. Today, web designers are ubiquitous across the economy. Similarly, we may see prompt engineers and AI specialists become a new profession in the services economy.

Ultimately, the panel agreed that those facing the biggest risk and disruption are lawyers and legal professionals who do not adopt generative AI in their workflows. They will see their competition become more efficient while they remain loyal to inefficient ways of doing things and conducting business.

The lawyers who don’t embrace this, they’re the ones who are at risk. Those who see it as an opportunity and the ability to continue to morph and change into the role of the future, that’s fantastic.

Danielle Costin, Projects and Innovation Director at Shoosmiths

Ready to innovate and embrace the future? Request a meeting to learn about Everlaw’s award-winning AI-powered litigation platform here.