Artificial Lawyer recently caught up with Emily Foges, the CEO of UK-based legal AI company, Luminance, which is backed by law firm Slaughter and May and Invoke Capital and launched in September last year.
We discussed how Luminance is doing less than nine months since its official launch, how Luminance is ‘language agnostic’ and where the company is heading next.
Emily, along with many other great speakers and experts, will be appearing at the Future Lawyer Summit on 4 May in central London. (Yours truly will be chairing the session on legal AI).
If after reading this interview you’d like to hear more about legal AI and other areas of legal innovation, Artificial Lawyer readers can get a special discount of 25% for the Summit. Simply enter ‘AL25’ when purchasing your tickets.
But, now, back to the interview…
Emily Foges first explains how she came to run one of the UK’s handful of legal AI companies. Unlike in some cases with legal AI bosses she is not a machine learning techie, nor a ‘recovering lawyer’.
She says that the move into legal tech came about by good fortune. After working in the field of large corporates and M&A for around 20 years she had wanted to try a new challenge and to work at a start-up. A friend mentioned a potential role at Luminance and the rest is history.
Foges’s previous career included stints at telecoms giant BT and financial data company Equifax, and nearly six years at UK betting company Betfair, where she managed, among other things, the integration of two strategic acquisitions into the Betfair group: TVG in California and Pokerchamps in Copenhagen.
In short, Foges knows a thing or two about the world of corporations and also mergers. And that’s handy when the product you are selling is designed to help lawyers working on mergers and acquisitions. It also helps if the plan is to grow a start-up into a far larger concern.
Foges joined the Luminance team in May last year ahead of the official September 2016 launch and some time after the company’s brain trust up in Cambridge had worked with former Autonomy founder and Invoke Capital boss, Mike Lynch, to get the first version of the AI due diligence software up and running.
Over 2016, UK-based M&A specialists, Slaughter and May, also joined the party, helping to shape, test and iterate the due diligence analysis software, ahead of its public launch.
Now, in April 2017, Luminance has grown rapidly and just moved into new offices on the Strand in London, which are ‘closer to lawyer-land’ Foges observes, as compared to the rather posh offices of Invoke further West on Pall Mall, just up from St James’s Palace.
After what is less than nine months in full operation Luminance is doing well. In fact, Foges says that they were recently looking at their early sales targets and had now clearly gone well beyond them.
In addition to Slaughter and May on the client list, there is Houthoff Buruma, the leading Dutch firm and two Scandinavian law firms, that have all publicly announced long term deals with the legal AI company.
However, Foges points out that is not the whole story. The total number of paid up clients is actually 12, with at least a dozen more in pilot stage. Among these are several inhouse legal teams at corporates and, perhaps what for some people may be a surprise, some LPOs, which are also piloting the system.
Getting potential clients signed up for pilots is perhaps a little easier for Luminance as it doesn’t charge for testing its technology, as some legal AI companies do. Though, that said, some smaller start-ups have no choice but to charge for pilots as they need the revenue to fund their operations.
Foges explains with understandable pride that the client list is now global, more so than the company had first expected. Clients range from US law firms to firms right across Europe and into Asia and Australia, including a firm in Singapore.
Not all firms want to go public and some are ‘shy’ about publicising their use of AI, she notes. Other firms are more than happy to publicise the relationship as it brings kudos. Either way, the use of Luminance and legal AI in general is spreading at a steady rate across the world’s leading law firms.
If we add in the successes of competitors such as RAVN and Kira, and consider the contract sign ups by AI-powered legal research companies such as ROSS, and the sign ups to AI-driven litigation analysis systems such as Ravel and Lex Machina, then we are looking at a legal world that is truly in full adoption mode of legal AI.
Mind Your Language
While several of Luminance’s clients are primarily using its natural language processing (NLP) due diligence capability for English language legal documents, Foges points out that the technology is ‘language agnostic’.
At first glance that may seem paradoxical, i.e. how can a system that is based on the understanding of written text not be impacted by a change in language, especially as Luminance, as with many other NLP systems, trained itself initially on English? How can the system spot specific legal clauses in Dutch, or Norwegian, for example, unless it has been built from the ground up to handle these languages?
The reason is that language itself is not the issue. It’s the patterns in the language that matter to the AI technology.
‘Luminance pulls out the patterns from the legal documents and these are assigned a tag by the lawyers,’ Foges explains.
That is to say, the NLP and the neural network system it uses is able to discern patterns of legal clauses in the texts and then isolate them no matter what language they are written in.
All languages implicitly structure certain phrases and ways of saying something in specific ways. In the legalese of each language this pattern is often even more distinct. If those patterns can be spotted and tagged then this approach to legal AI is potentially universal and could be used in any language that has a cohesive structure.
But, surely something like Mandarin would be a problem? Foges says the team have discussed this, and that in theory at least, there should not be a fundamental problem even if dealing with ideograms is different to a Latin, phonetic script.
‘It’s still a pattern,’ Foges says, referring to looking for a certain clause type in a legal document written in Mandarin. And, if it’s a pattern then it can be found and identified. This is the triumph of AI over Big Legal Data, one might say.
Moreover, this perhaps underlines the huge potential legal AI companies of this type may have. I.e. Luminance does not just have a chance to tap into the legal needs of firms working in English, but the world’s legal market. Given that the global legal market is estimated to be $700 billion in value, then even if only a very small fraction of this were to be considered as legal AI document analysis territory, then we are still looking at a future multi-billion dollar industry in the making.
Foges clearly has enough to keep her busy, but due diligence is just the beginning, she says. The company had to start somewhere and M&A diligence was an obvious starting place given the potential demand. But, this is not the end of the story.
‘We started off with due diligence in part to have a clear message. But partners in the firms we work with are interested in seeing how Luminance can be used in other areas. We can see applications of our technology to areas such as litigation,’ Foges explains.
‘Compliance is another area we are looking at,’ she adds, noting that Brexit and all the legal and regulatory changes that will come with it potentially create a greater demand for compliance projects that are high volume in terms of data and therefore are perfect territory for legal AI. Changes to rules over how data is held, from the GDPR changes in Europe to PCI credit data rules in the US, all create other opportunities.
Then there is working more closely with inhouse legal teams. ‘GCs say they don’t know what data they have got. They want to know how to get their arms around it,’ Foges adds.
Banks in particular are drowning in data and documents, in part because as technology advances the total quantity of information large institutions create multiplies. Organisations such as banks then need to try and manage and control this mass of information. Legal AI can come to the rescue here as well, and perhaps just in time.
As noted above, Luminance is also piloting with LPOs, a step that at first may seem to cannibalise their often manual business model. But, presumably the LPOs can see that they need to adopt legal AI tech, rather than pretend it’s not there. The clients are certainly aware of it now. So, perhaps there is no choice? After all, if you can’t beat then, join them, so the saying goes.
All in all these are interesting times for Luminance, which is having a greater impact on the legal market that can at first be seen if based only on publicly announced clients.
Foges’ optimism and success during what are still very early days for legal AI also suggests that much more is to come. One might say that if a single legal AI company can get this far in seven months, then the overall impact of the entire legal AI sector over the next few years is likely to be a game changer.