Solomonic talks data

Litigation data and analytics startup Solomonic was recently rolled out by Herbert Smith Freehills across its UK disputes team and they are working closely with Stewarts Law. Don’t know much about Solomonic yet? You do now.

Who are you and where you do come from?

We are Solomonic, a UK based and focused litigation data and analytics start-up. Started in late 2017 we have our first commercial offering, analytics and data covering the past 15 years of the Commercial Court, live in the market. We bring together legal experts, (our founders Gideon Cohen, a barrister at One Essex Court, and David Cohen, a tax solicitor with Keystone) and data science (co-founder Dr Henry Stott, a behavioural data scientist who also runs professional services consultancy, Decision Technology). Our chair is experienced M&A strategist Natalina Bertolli and we have Chris Style QC and Ian Terry as our advisors.

Can you tell us a bit about your tech stack?

To build our minimal viable product (MVP) we quite quickly jettisoned any AI or other seemingly fashionable technology and used very smart trainees and barristers to triple review every judgment. That rigorous data does have some sophisticated multi-variate predictive models thrown at it, but in the main we use tried and tested robust statistical modelling to drive a dashboard which contains a series of data views that users can filter against by legal subject, judge, compensation sought, application types and so on.

As we move beyond MVP into fast growth we have gone back to looking at and investing in machine learning tech to make our data mining, extraction and ingest more efficient, without removing the need for experienced barristers to review each judgment.

Ultimately, the best results and outcomes for clients will come through the intelligent combination of the litigators’ expertise and experience and our analytics.

There is a reality check here, which may not be that exciting to those believing AI is going to take over the world, but the good news (for lawyers anyway) is that the technology of today is not good enough to replicate their analytical, interpretive and decision-making skills.

We know that in the high stakes and high value world of litigation, our analysis has to be rigorous to be trusted so we can’t afford to produce sloppy data.

The good news is that Innovate UK shares our vision for augmenting human intelligence and has awarded us funding to optimise the role AI can play in the expansion and acceleration of our business through machine learning and neuro-linguistic programming (NLP) and we will be working with our partner, Warwick Business School, to do just that over the next 12 months.

And the clients you work with?

We have been extremely lucky to have developed some great partnerships, especially with Stewarts, who are a very smart and innovative firm and with Herbert Smith Freehills, who have a really sophisticated vision for the way in which litigation and risk advice is designed and delivered.

Right now, with our focus on the Commercial Court, our market is mainly with the leading UK firms, but we are seeing interest from US firms and from litigation funders. Our next area is the Chancery Division – we have already started on that, planned for release in May and then tax and employment. Thereafter we are going to be led by where the market demand takes us – that may be international.

What problem is it that you are trying to solve?

We help solve two problems. The first is around helping litigators and their clients better evaluate risks, probabilities and therefore decision-making related to a piece of litigation. Today so many of these decisions are made without access to rich data, especially valuable where the outcome slides towards the 50/50 mark. This is especially painful for clients in data-driven businesses and particularly at board level.

The second problem we help solve is the litigation research task. Our data and analytics are focused on outcomes and variables that impact your strategy and prospects – which claims and applications win, which don’t, what impact the judge has, which barrister you should use and which experts you should select. All our data resolves to the underlying case law so it makes it very quick and easy to get the right answers. In addition, because we have tried so hard to be comprehensive, you are less likely to miss those needles-in-a-haystack cases because we have mapped them, whether by outcome, by judge or by case type.

Can you give us a cool example or two of your tech in action?  We had one a couple of weeks ago. Our data shows in nice and very statistically robust terms that Section 68 arbitration appeals fail over 85% of the time. The litigator knew the prospects were poor, but to have the data to hand (all S68 appeals for the last 10 years) in a simple–to-grasp set of pie and bar charts made it that much easier to persuade a client that appeal was not really an option.

What do people assume is possible that isn’t? 95% outcome prediction of a piece of commercial litigation, in a world where cases that go to judgment take on average over three years. No sane litigator or client believes that’s possible.

Where do you see this going in terms of future roadmap? Well, we want to provide a comprehensive commercial litigation view by covering Commercial and Chancery, then areas with high value litigation, tax, employment and potentially others. The Appeal court has come up and we also think once we have our operating rhythm right, we can focus on other jurisdictions so users can compare prospects.

What are the challenges? Data, data, data. Data in the UK is so hard to aggregate, mine and ingest because it is so piecemeal and patchy in how well it is assembled, recorded and stored.

How do we address the lack of publicly available case law?  By getting HMCTS to focus on the basics, not just basic document provision, but all the basics. It feels like we have been seduced by online courts and dispute resolution where, right now, what everyone wants is for the basics to be done better. That and a couple of court document API’s would be nice.

This interview first appeared in our free monthly newsletter, the Orange Rag – sign up here: