Technology could level the playing field

When it comes to artificial intelligence (AI) for law firms, we are currently at the ‘frothy end’ of the hype cycle, as Richard Kemp, founder of Kemp IT Law, describes it.

Despite the noisy proclamations of software providers and excited articles from relevant experts, law firms’ actual adoption of this innovative technology has been rather cautious.

But according to the Law Society’s Lawtech Adoption Research report published in February, while lawtech has a long way to go before it matches the success of fintech - which is currently transforming the financial services industry - the market for legal technology has been gaining momentum over the past 12 to 18 months. 

As one market commentator quoted in the report put it: “Lawtech has become somewhat ‘cool’, a bit like fintech but a less sexy cousin. There has been a huge ramp up in awareness and press exposure over the last two years, and the appetite for lawtech iws definitely maturing.” 

AI technology, which includes ‘machine learning’ tech, has been springing up in all areas of legal practice. It is well-established in the e-disclosure arena, and these days the courts positively expect parties to use technology to reduce disclosure costs. Now it is reaching further, into legal research, corporate due diligence, contract automation, e-billing, and even litigation itself. 

Peter Wallqvist, VP of strategy at software provider iManage, explains: “Machine learning is very applicable for tasks that are very easily definable, and tend to be very arduous for a person to do. It is for the sort of thing where you could instruct a non-qualified person to do it, and it would only take a few sentences to explain what needed to be done.”

He adds: “Say in real estate work, there might be 1,000 leases involved, and you don’t want to sift through every one to see the policy and monthly rent. Now you can have a machine do the bulk of the work. 

“You simply point to what you want a few times, and the machine will learn and pick out the information on its own. It won’t be possible every time; it depends how varied the task is.”

With the ability to cope with mountainous data sets, one could be forgiven for assuming this kind of tech is mainly aimed at the big legal giants. But not according to Wallqvist.

“One of the reasons why larger firms outcompete smaller ones is because they can snap their fingers and get a room full of paralegals at their disposal,” he says. 

“But the interesting thing about this technology is that it works just as well for smaller firms as for bigger practices. It’s an incredible equaliser. It removes the competitive advantage of being able to start throwing bodies at a task.” 

After all, the cost of this tech will usually be paid by the client and can be expected to be lower than the billable hours stacked up by paralegals. But AI tech will not be suitable for all firms. 

“Where we deal with law firms below the top 30, the ones where this sort of technology is most successful tend to be those that are pretty specialised,” Wallqvist notes. 

“If they do [a process] hundreds of thousands of times a year, the return on investment will be much greater. It tends to be firms that deal in personal injury, insurance claims, conveyancing – areas where they have had to become much more efficient at what they do to survive. But that experience is slowly creeping up to encapsulate more work that was previously considered to be bespoke”. 

Drew Winlaw, COO at Wavelength Law, adds that machine-learning technology can be particularly useful for niche firms. For example, he works with one firm that specialises in planning – a field packed with potentially useful datasets. 

That includes data from the planning inspectorate; data from the planning court, which might be two to five cases a week; and detailed data on the planning obligations imposed on developers to build roads, playgrounds and so forth as a condition of allowing the development.

“Being able to draw insights from these kinds of data sets is going to be very interesting for the future,” observes Winlaw. “Increasingly, clients want answers to the bigger questions, such as, ‘what are the chances of me running this argument and getting approval from the authorities’? With access to enough data, a law firm can gain ‘experience’ of a particular area more quickly, by drawing insights from what has happened before”. 

Wallqvist adds that what clients are looking for from their lawyers boils down to ‘knowledge’; and the technology helps lawyers to enhance that knowledge. 

That applies not just to external data sets, but also to the exploitation of information already contained within the firm itself, which is often not as accessible as it should be.

But while the technology is advancing all the time, for Kemp, there is one tasty nut that has not yet truly been cracked: due diligence in mergers and acquisitions work. “Due diligence currently takes up around 40 per cent of a law firm’s fees in a typical M&A deal,” he says. 

“But there hasn’t yet emerged a machine-learning tool that can understand the contracts that form part of due diligence, to extract the information automatically. There has been huge progress, but there is more to come”.

The solicitor notes that when this breakthrough finally comes, it will not necessarily be law firms leading the way. It could be the Big Four accountants; an Indian outsourcer; or some chap in a warehouse that we haven’t even heard of yet. 

The big law firms might be inclined to try and cook up their own AI recipe using their special secret sauce; but the corporate clients themselves will be pushing for more of a ‘platform’ approach, whereby the tech will be more widely available. 

“It’s going to be a very big market,” predicts Kemp, “and the platform approach is likely to be a benefit for the smaller law firms. Within a few years it will be possible for people to buy in these AI services from the cloud, even on a pay-as-you-go basis.”

Kemp is enthusiastic about the potential of AI tech for law firms, but he has a warning. “It is easy to see machine learning as a panacea; a solution in search of a problem. It all comes down to being clear about what the actual problem is that you want to solve.

“If you invest in the technology, what are you going to do with it - increase prices, increase throughput? How are you going to make more fees from it?”

According to Winlaw, “law firms tend to be thinking about the individual client, rather than the wider market need.” But they need to be bolder. 

“They need to ask themselves, if we did crack this problem, what is the potential of it? How would we cope with scaling it up, and in which geography? It is the ‘thinking big’ element that is not very commonplace in professional services environment,” he remarks.

Litigation weaponry

Another area ripe for the use of AI tech is litigation. The relentless tide of judgments emanating from the courts form a deep pool of information. 

Once this has been sorted through and tagged, the AI technology can allow this vast body of data to be analysed in any number of ways. 

It can tell you what percentage of a certain type of claim have succeeded in the past; which barristers have the best success rate in the field; which experts have been complimented or criticised in judgments; how often a particular defendant tends to settle cases; the track record of a law firm in a certain type of case; or which other cases have dealt with similar arguments.

Edward Bird is chief revenue officer at litigation analytics start-up Solomonic, whose software is currently being rolled out by Herbert Smith Freehills, and covers most Commercial Court litigation from the past 15 years.

He explains how the technology is not only useful for assessing the initial prospects of a claim – and managing the client’s expectations in that regard – but also for informing the ongoing strategy of the litigation as it develops. 

One interesting aspect is the way the tech can analyse how much of a risk a particular judge might be. 

Bird explains: “Say you had a claim about the sale of a mine, and there was a sum of money that would only need to be paid over to the seller if the buyer obtained a licence, or a particular loan facility in relation to the mine. 

And so instead of that particular loan, the buyer obtained another form of investment instead [to avoid triggering the payment clause in the contract]. That is the type of claim where it would need a commercial interpretation of the contract in order for the seller’s claim to succeed.” 

Bird adds that the software can point to various elements that might indicate whether a judge will be inclined to take a rigid interpretation of contract wording, or a more flexible approach. 

For example, the tool can show how often the judge has used terms such as ‘commercial sense’, ‘plain’ or ‘clear’; how often they have mentioned or agreed with a certain authority; or their propensity to discuss precedent. 

Put together, this can help lawyers to build a picture of what approach a particular judge is likely to take. 

The technology is not intended to replace lawyers, but could provide another weapon in their armoury. And while it is easy to see why firms that litigate regularly might want to make use of this kind of technology, for those firms that take cases to court only rarely, a quick dip into a resource such as this could be even more beneficial. 

In the future, one could expect this type of data mining technology to become available across all courts, even down to the decisions of cantankerous district judges in far-flung jurisdictions. 

Breaking down barriers 

Given the potential benefits of AI, why has the legal sector so far been relatively slow to embrace it? Legal technology expert Chrissie Lightfoot says: “There are missed opportunities that firms would certainly find genuinely useful – for example, many firms are not even using chatbots yet, for early stage assessment and form filling. 

“But law firms can be quite overwhelmed by the legal technology companies bombarding them all the time; they are trying to work out which tech is actually useful to them.” 

Lightfoot adds that firms can also struggle to get buy-in from partners to invest in technology, particularly where they fear that the technology may actually reduce the amount of time they can bill for work. 

And even where firms do sign on the dotted line and purchase the tech, it is not always used as extensively as it should be in practice. 

The Law Society’s research report identified that the ‘internal sell’ of new technology was the hardest part of adopting it, with some firms admitting to ‘sporadic and inconsistent’ implementation.

Winlaw says: “Sometimes law firms spend all their energy building the case for spending the money on the tool, and underestimate the effort of using the tool – the time that will need to be spent on lawyers learning how to use it, developing protocols for using it safely, and the [money needed] to keep it working. There are ongoing overheads.”

He adds that software providers could focus more on sitting alongside firms after the purchase has been made, to help them get the most out of the technology they have purchased.

But the signs are that the legal sector is about to begin embracing AI technology in earnest. As Lightfoot puts it: “So far, the hype has not been matched by the scale of actual adoption. 

But in the next three to four years, the implementation curve will be very much on the up.” Watch this space.

Rachel Rothwell is a freelance journalist