Due diligence is a key area for commercial law firms and one that many legal AI companies have targeted. How does global firm Linklaters approach this vital element of its work for clients? Artificial Lawyer spoke to Greg Baker to find out.
London-based Baker recently has taken on a new senior lawyer role focused on practice transformation (AI and innovation) within the global corporate practice at Linklaters. Central to this role is building a team for tech-supported due diligence, in effect a ‘centre of excellence’.
His role will not be fee-earning and will build on the long-term strategy of developing ‘innovation lawyers’ embedded within all of the firm’s main practice groups.
In terms of that work, ‘some of it will be AI’ Baker said, and it will be closely connected to the Linklaters-developed tool ReportiQ, which was launched out of pilot in 2023 and is a collaboration with the US vendor WizDocs.
The driving force behind ReportiQ, alongside Baker, have been Timo Engelhardt (Munich M&A partner) and Jennifer Klement (Product Owner). The end-to-end solution, helps with ‘changes in the Virtual Data Room, allocating documents for review, collecting questions and managing the Q&A process, offering a vetted library of suggested language, and producing the final report with a single click’.
Clients also get a ‘unique risk profile visualisation, [and] easy information access, as well as progress monitoring options’. So far €35bn+ worth of deals have been supported through the platform.
And while this has been a success, Baker explained that the broader approach is to think clearly about ‘the process and the toolkit’ and then look at ‘buy, borrow, or build’ in terms of where the core tech of a solution comes from.
After that, it’s about thinking further about the wider process the solution will be part of, where automation makes sense, and where standards can apply.
GenAI and the Bigger DD Picture
Meanwhile, genAI has added a new dimension to due diligence.
‘We have [also] used Kira, for example, and that has a machine learning approach, while with genAI it’s less about training [of the AI] and more about prompts,’ he said.
GenAI finds its way into improving the overall workflows, Baker noted, then explained: ‘The buy side in a deal will have lots of Q&A [in relation to the document set]. This can be very painful.’
I.e. other parties are sending in dozens of complex questions every week as the deal progresses. By using genAI – which is very good at language understanding – they can group questions or combine them, and also remove duplicates.
‘This can provide a quality of life improvement,’ Baker noted.
Then there are other uses that help the lawyers with some of the volume end of things, such as zipping through hundreds of documents to count up how many leases a client has, for example. That’s not super-complex, or high value work, but it’s necessary and would take a manual-only team a very long time.
Baker stressed that ‘due diligence is not just issue spotting’, and there is much more that AI and other tech can do to help.
That said, there are still some crucial aspects that do relate to issue spotting – although perhaps not as simple and clear as finding a thorny clause that deviates from the norm.
‘Due diligence for a private equity deal can involve the client wanting to know if there is anything that may harm the value of their investment,’ he noted, which is not always as simple as dragging out a clause and its terms, but may involve interpreting the meaning of multiple sections of text across a document set.
‘There is a lot of nuance,’ Baker added. ‘And genAI allows us to be more nuanced than machine learning tools that need a lot of upfront training.’
In terms of accuracy, Baker underlined that this is of course super-critical, but there are times when the goal is ‘to look for where you need to steer the lawyers’, so they can then dig into the issue, rather than it being a matter of extracting specific text. I.e. sometimes it can be like using a ‘targeting computer’ that improves the aim.
Baker also underlined that due diligence is an area with broad applications, and goes beyond M&A deals into real estate and finance transactions.
The Benefits
So where does that get us? It’s clearly not primarily about replacing lawyers in the diligence process, but rather ‘it informs the negotiation of the deal’ – which is a good way to look at things, i.e. the AI tools are there to elevate the work of the lawyers, so they can complete better deals for the clients. In short, create value, as well as shave away inefficiency from laborious – albeit essential – work.
Given all the above, how much of an M&A deal can tech help with, at least when it comes to diligence? Baker estimates it can range from 10% to 50% of the total time involved in a matter, so it’s clearly significant.
Naturally, much depends on each deal, with due diligence including the main doc review process, plus other workstream factors, which include Q&A, presentations, drafting the report, and interim reports.
Moreover, on top of due diligence, Baker noted that ‘our matter blueprints and process mapping identify a wealth of other opportunities for tech to support, from AI-supported NDA review in auction processes, ‘what’s market’ analysis in the negotiation, through to preparing issues lists and project management assistance’.
Overall, what we see here is a mature approach to the use of AI technology and thinking clearly about workflows. It’s crushing process work – yes; but also then the goal is to apply the results of this to improve the bit that really matters to the clients: cementing a deal in their favour. In short, AI is a means to an end, not an end in itself.