What Happens if We Network Together Legal AI Systems?

Will there ever be the equivalent of a ‘Microsoft Office suite’ of legal AI and other advanced legal tech software that a law firm, or inhouse legal team, will be able to buy as a package?

Could you put together a range of this new wave of legal tech that contains a wide collection of AI systems, along with other advanced forms of legal technology, such as process automation, legal bots and expert systems? Could it all be integrated to make one single, super powerful, super useful legal AI ‘suite’?

Software’s Fuzzy Edges

First, let’s take a short walk down memory lane and consider the birth of word processing systems. Today Microsoft Word has obliterated the competition, but it was not always that way.

Artificial Lawyer is not a software archaeologist, but luckily someone invented Wikipedia, which has also come to dominate the open source information world. What is clear is that there was a steady evolution in word processing software, with pioneers such as IBM, Xerox, Canon and Hewlett Packard, back in the 1970s and early 1980s.

As the number of business-focused word processor companies multiplied and fought for market share a new wave of companies also appeared, such as Microsoft and Apple, which were focused on the emerging desktop market. There were many others too, but they are gone now.

Microsoft launched the first version of Word in 1983, thirty-four years ago. Then came Excel in 1985, and then PowerPoint in 1990. In 1993 the US company launched Exchange Server as an email system. And in turn when you open Outlook 365 today you see a myriad of additional types of application available, ranging from Skype for video calls, to Yammer and Delve for collaborative working.

The software has not just evolved, it has networked together. It has created a synergistic whole that is far more powerful than the sum of its parts.

But, why did this happen? The first point has to be that one reason why Microsoft was so successful, was not just that its software was easy to use and didn’t need a computer science degree to figure out, but because the company recognised people’s technology needs were not discrete.

That is to say, although there are many different tools for specific tasks and always will be, that doesn’t mean they cannot be linked together.

People’s technology needs have fuzzy edges. The text of a Word doc needs to go into a PowerPoint presentation, some of the data in that needs to be stripped out and put in Excel. Then you want to email the presentation, if you still even use email, and perhaps instead you prefer a collaborative platform such as Slack. And so on. It all links together. Human demands outstrip any single discrete use case of a tool. We want it all. In one place. Easily accessible.

In short, technology seems to have two contradictory forces at play. One force is driving innovation of new tools, exploring new niches and building upon past IP to venture down a particular path. The other force is constantly trying to wed all these niches together to build a networked whole that people can more easily use.

Networking Legal AI Systems 

Imagine a scenario where you have a suite of legal AI tools. Some are more discrete than others. Some need considerable training on available data before they can be let loose to scour and explore documents, others have such universal uses they can just ‘get to it’ with little or no training, i.e. their natural language processing (NLP) recipes are broadly applicable to a wide range of types of legal texts and document formats.

In fact, one could argue that some legal AI companies have already started to move in that direction, such as RAVN, with its broad offering of different AI tools, going from due diligence doc review to sorting between privileged and non-privileged documents. The underlying tech is similar, the tool’s end use is different. Meanwhile, Kira seeks to provide a sort of ‘universal adaptor’ approach, permitting clients to create whatever they want in terms of doc analysis tools from the core NLP technology the company provides.

But, what if we take this further? What if we add to such doc review tools some legal research systems, whether like ROSS, which is practice focused, or companies such as Premonition, Casetext, LexMachina and Ravel, which focus on case law and litigation analytics? And then perhaps let’s add in some interactive interfaces, such as Neota’s and Rainbird’s expert systems that help users reach an answer to a specific legal query?

And if we are going that far then we may as well add in the capacity to quickly and easily build and deploy your own legal bots. And to finish off, let’s also bring in the new wave of process automation technology being pioneered by companies such as Autto.io.

This suite would naturally link to the existing document automation software on the market, as well as current CRM and KM systems that law firms use.

Imagine all of that together on a lawyer’s desktop, accessible via a couple of icons and operating with an easy to use interface that required little or no technical knowledge of the software that powered it.

How much would that increase productivity in law firms? How easy would it be for a partner and their team to create a wide-ranging legal AI ‘product’ of their own for a client that added value? The leap forward would be incredible.

Barriers to Future Integration

Much as the idea of a ‘law firm OS’ with its bundled ‘legal AI suite’ coming to every lawyers’ laptop and smartphone on the planet may one day be a reality, the barriers to getting there are considerable.

The first and most apparent point is that for now legal AI, and those technologies such as expert systems and bots, are being developed by a myriad of teams all over the world. From Toronto, to New York and Silicon Valley, to London, to Paris, Amsterdam and Berlin, to Sydney and Auckland, there are literally dozens of New Wave legal tech companies in action.

At present, collaboration between these teams is minimal, with interactions of HighQ with RAVN and Neota, among those few that have been publicised, though these have been primarily ‘friendly’ joint projects, rather than attempts at business consolidation.

Putting on a strategy consultant hat for a moment, the reason why start-ups, as well as some of the more developed new legal tech companies, tend not to want to integrate fully with others is that the founders still believe their company will be the one that establishes itself and gains significant market share.

Why dilute the brand, the offering and perhaps most importantly the ownership when you could be the new tech company that one day leads the market?

In which case, is there a business case for integrating and combining a wide range of legal AI and other legal tech? For sure and a very compelling one. There is only so much money that law firms will spend on legal tech, even if their budgets are growing. Market share is key to any sector. Hence, a suite of different legal AI systems and other advanced legal tech would be an awesome value proposition that could swallow up market share.

But, if a company is still trying to prove its concept, build a revenue stream, satisfy early external investors and secure a reliable client base, then it is simply too soon for this kind of large-scale consolidation to take place.

Moreover, if you have a legal AI company, with its own founders, its own investors and its nascent client base that in some cases may partly be in pilot mode, then the drive to combine with another company, with different founders, different IP, different investors and its own nascent client base, is probably not strong enough to succeed, at least in these early stages of company growth.

In many cases the legal AI companies may see each other as competitors fighting over exactly the same potential clients. Or perhaps they plan to move into a rival’s niche in the future and would rather wait to develop their own tech for that rather than seek a merger. Moreover, the investors may not want to give up or dilute their share of the business before the company has grown sufficiently. And given that legal AI is so new still and total revenue streams for most pioneers remains narrow, then most investors will no doubt want to wait it out.

While, if the goal is to sell to a larger and established company, the same points may hold true. I.e. why sell now even before the market has fully tuned into what legal AI can do? To do so needs a compelling business case, though this has happened.

The November 2015 deal between Lex Machina and legal publishing giant LexisNexis made a lot of sense at the time because the case law analysis system needed tons of legal data to make itself a valuable tool. LexisNexis also clearly saw a business use for adding AI-led case analysis to the search and analysis systems it had already.

The Future

Will we ever get to a fully integrated, very broad suite of legal AI technologies and other advanced legal tech, all packaged up into a single offering that you renew every year on subscription? I think, yes, we will. But not for a long time to come.

We may perhaps see two or three major legal AI suites of this type emerge, with perhaps two or three other smaller suites of New Wave legal tech that focus on such a specific part of the legal market that they are able to form standalone groups of software, just as today we have Adobe and its mass of design software standing outside of the other aggregated software suites.

But, for now, what does all this mean? It means that the legal AI market is in for a very long and interesting journey and the future is still very far from mapped out or belongs to any particular company or type of AI technology.

And that’s exciting because it means there is still everything to play for. Watch this space.

This article was written by Richard Tromans, Founder, TromansConsulting.