Laura Bygrave and Adrian Rainey at Taylor Wessing on adding value with Kira Systems
Taylor Wessing set Kira Systems to work on identifying trends in venture capital deal terms, providing its clients with even greater insight into a fast-moving market. But that’s just the beginning, say Laura Bygrave and Adrian Rainey.
Kira Systems has established itself as leading the way in helping law firms to make themselves more competitive and efficient with the help of artificial intelligence (AI). It is perhaps best known for the ability to step into the traditional due diligence process – automatically identifying and assembling key information, and so releasing lawyers to focus their time and attention on the more complex aspects of legal work.
However, the innovation team at Taylor Wessing had something else in mind when they were assessing the market for machine-learning solutions to support the firm’s contract analysis work.
In 2019, even leading law firms can no longer rely on simply completing pieces of legal work for their clients; clients expect their advisers to provide additional data surrounding the total body of work to continuously help those clients sharpen their strategies. Taylor Wessing realised that the algorithms of Kira held the potential to provide its clients with an additional layer of very valuable market insight.
Head of the corporate technology group, Adrian Rainey – whose practice advises tech companies on venture capital fundraising through to acquisition or sale – explains: “In negotiations, lawyers will often find themselves in discussions about how ‘typical’ a characteristic of a deal may be. It’s very helpful to be able to indicate that ‘X% of 200 previous financings in 2018 included Y’, for example.”
Just like the due diligence work itself, accurately pulling that data presents a very laborious task for the time-poor lawyer.
What’s the use case?
Rainey continues: “This is an unusual use case in that we needed to train the core product entirely new concepts.” Kira arrives already primed to automatically identify and extract over 600 built-in data points. However, the initial 20 provisions Taylor Wessing wanted were not among these. “We really needed to train it on a niche within a niche – emerging tech companies in the UK,” he says.
The fact Kira could be engaged in such a way was the critical factor in its selection as the tool for the job, adds innovation manager Laura Bygrave. “One principle of the firm’s technology innovation strategy is not to focus on a single use case, but to consider how we can deploy AI across multiple practices in the UK, and then internationally as well.” Flexibility and speed to market are key drivers.
However intelligent your machine, of course it still initially relies on a subject-matter expert transferring their human knowledge, and thereafter keeping that intelligence up to date and governed. Bygrave’s innovation team brought together 30 corporate partners and associates from the firm’s London and Cambridge offices for a series of “relays”, consisting of consecutive two-hour sprints to train Kira up to an appropriate level of precision within a couple of days.
An oft-heard innovation mantra is to ‘fail fast’ – something that clearly applied. Bygrave says: “We decided it would be best to experiment very quickly, to reach a point of knowing whether to invest further lawyer time, or to walk away and look elsewhere.”
Winning the time of partners for an exercise like this is also quite rare, she adds. “It was therefore vital for us to prove that the product worked for the use case in principle,” she says. “We needed buy-in from right at the top, right from the start.”
Rainey adds: “The truth is we gave a lot of thought to whether we ought to involve partners. However, as Kira needs to be able to spot very specific variations within terms, we felt that omitting partner experience would result in an inferior output.”
The system will now be piloted in a similar fashion across the firm’s private equity, banking and finance, employment and real estate teams, to put the business case for scaling beyond all doubt.
“Ultimately, new technology has to be embedded as a business-as-usual process,” says Bygrave. “For that, it needs to address a genuine pain point to get people going and staying on the change journey.”
Spot the picture
An unexpected bonus was that the lawyer-relay process lifted the lid on a second use case for the firm.
Rainey says: “If Kira could recognise all of these specific technology deal terms, we realised it could also effectively review a set of drafts – for example, from the other side of a deal – producing a summary, or highlighting the changes.
"In the short to medium term, a lawyer will definitely review as well. But an automated snapshot of 80–90% of the full picture could be really useful. That’s a win we weren’t anticipating.”
Bygrave picks up the baton. “We were able to produce high-level insights that could be sent to clients in minutes, leaving lawyers to focus on the more complicated deal points.”
This is only one of many innovation initiatives underway at the firm involving making the most of machine learning. The TW ‘TechSet’ consists of eight products, which combine legal experience with technology to provide advanced methods of client delivery. This includes the recently piloted TW:detect, also helmed by Bygrave, which was developed in response to growth in malware campaigns and which scans clients’ websites for potential compromise.
Meanwhile, the future for Kira with Taylor Wessing innovation is bright … if full of a lot more training. But this is one solution that has shown itself worth the investment.