A new legal tech startup, curiously named LegalAI, hopes to automate the pre-court litigation process for consumer claims case assessment. The company also hopes to work closely with insurers to roll out the product.
It’s the brainchild of Germany-based Tessia Tober and Arpit Bajpai (pictured above), and works like this:
- Someone has a consumer claim, e.g. a problem with the hotel during a package holiday.
- They come to LegalAI and fill in an online questionnaire.
- Using NLP analysis to match previous claims of a similar nature with the new dispute, LegalAI then provides an automated case assessment.
- The assessment can then be used to make a claim against a company, for example, a package holiday provider.
- LegalAI then operates on a no-win, no-fee basis and claims a percentage of any compensation the consumer wins.
It’s not the first time prior case analysis has been used to try and predict future outcomes. CourtQuant, for example, also tried to do this. And it’s not the first to try and galvanise consumer claims using a tech platform, for example DoNotPay has pioneered the way here.
But, from what Artificial Lawyer can see, LegalAI is spanning the gap between these two companies. CourtQuant was looking at fairly complex cases, which even for medium-size disputes had many disparate moving parts. Meanwhile, DoNotPay seeks to help people by providing a pathway through an already well-developed consumer claims system, e.g. refunds for late flights. Their model is not to try and do case assessment, but rather to use a digital template to help a consumer input the data needed to access compensation – and therefore is a lot simpler, (and perhaps that means it has a very high chance of working).
Artificial Lawyer spoke to the founders, Tober and Bajpai, about their project, which initially will focus on travel, tenancy claims, and traffic accidents.
‘In certain legal domains you can apply NLP to predict the outcome of a case. Most of the time this is a mater of classification. It’s a yes/no issue. Fundamentally, this is a regression problem as to whether someone will get their money back [for a holiday, for example],’ they explained.
Now, in Common Law jurisdictions this may be a little bit more complicated, but this startup is focused on Europe, and more specifically on German law, i.e. Civil Law, which has a ‘coded’ approach, one might say.
‘What we do is use a similarity engine to cluster cases together. That then gives us a similarity score. Then with doc automation we can make a legal document that contains legal reasoning, plus the ‘why’ and the ‘how much’ of the matter. That is then sent to the parties,’ they added.
All well and good. Of course, the challenge here is getting this right. At present they have trained their NLP models on legal data for 3,000 cases from a law firm that does consumer law, and they have also accessed public data.
That’s no doubt a good start and they are open about the fact they are in pre-Beta. Proving their models work well will be vital. They are also very open about the likely need to work with large insurance companies – and are already talking to at least one major insurer.
For the insurance companies this makes a lot of sense. They want to know what exposure they face, and more rapid case assessment would be a great benefit – as law firms such as Kennedys and its Kennedys IQ group have found, which has pioneered the area of case management for insurance companies with their Klaim platform.
Tober, who worked as a lawyer, concluded with the point that ‘there is a myth that all our work is complex. But [many] consumer law outcomes are simple and predictable’.
It all sounds like a useful endeavour, although Artificial Lawyer couldn’t help but note that ‘LegalAI’ was a bit of a massive brand name that represented an entire section of the legal tech world, and maybe a brand more closely linked to consumer legal claims was maybe more apt? The founders mentioned that the creation of sub-brands that related to different areas, e.g. claims for holidays, could be a future possibility.
At the moment they have some funding and are looking for additional cash to take their product on. And as mentioned they also hope to work closely with at least one major insurer. Good luck to them.