Automation of legal processes has not led to the revolution many were predicting, but generative AI is expected to change all that. In-House Perspective looks at what’s on the horizon and how it’ll affect legal teams.

Automation in the legal sector is nothing new. As far back as the 1950s lawyers were beginning to streamline business processes by using dictaphones to record their instructions on the go. By the 1970s photocopiers and word-processors had arrived.

The pace of change has accelerated exponentially in the past few years, however, with advances in technology meaning case management has been turned on its head. Machine learning has greatly cut down the length of time it takes to review vast piles of dense legal documents and everything from e-signing to contract automation has freed up lawyers to devote more time to high-value tasks.

Yet, while a 2016 report from global management consultancy Deloitte – Developing legal talent: Stepping into the future law firm – was correct to predict that ‘transformation of the profession is likely to be profound due to the quickening pace of technological developments, shifts in workforce demographics and the need to offer clients more value for money’, the point at which it predicted ‘the tipping point’ would occur was 2020.

Doil Son, Co-Chair of the IBA Technology Law Committee and a partner at Seoul firm Yulchon, says that so far automation has not had a substantial impact. ‘I’m sure that there will be, but at the moment AI [artificial intelligence] is not at the level that we and clients expect for professional services,’ he says.

If that is the case for private practice lawyers, most in-house legal functions are even further behind. Julian Hamblin, Vice-Chair of the Technology Law Committee and a partner at UK firm Trethowans, says lawyers in general have been slower than expected to fully embrace automation because of a general apathy towards trying out new technologies on the one hand and an in-built sense of caution on the other. ‘Most lawyers are broadly behind the curve in terms of adopting new technologies,’ he says. ‘From a private practice perspective, in most firms if you say to the partners that you’re going to roll out something like [an e-signature product] – it’s straightforward and everyone has to use it next week – they’d all go nuts and just say to their secretaries “you do it”. That’s how lawyers react – it was the same when the fax machine was introduced.’

But there’s a fear, he says, of making an error and being sued. ‘If you send a document out with a note saying “sign here” that’s pretty idiot proof,’ he explains, but if you send out a document requiring an e-signature and there’s an error, ‘you’ve got a bigger problem on your hands.’

Marc Hilber, Co-Chair of the IBA Technology Law Committee and a partner at Oppenhoff in Germany, says firms do use automation, just not for the kind of high-level tasks that were being predicted in the years leading up to the Covid-19 pandemic. ‘From a law firm’s perspective the question is what kind of work are you doing for the client,’ he says. ‘If you do high-profile complex work like complex contracts, automation is not really helpful and doesn’t really happen. That kind of relationship-based advice on complex matters remains the same, but for law firms that do a lot of that kind of work clients also expect AI-based services and automated services. What we are doing is working with workflow-based tools, for example for [non-disclosure agreements] NDAs or shareholder resolutions, to generate those standardised documents that are frequently required and are not really complex to draft.’

Behind the curve

In-house teams may be even further behind the curve than their private practice counterparts, but Tamar Krongrad, Networking Officer on the IBA Women Lawyers’ Committee and Head of Global Operations at alternative legal service provider LawFlex, says general counsel are looking at ways of incorporating such tools in their own functions as a means of adding another layer of cost control. ‘More and more we’re finding that in-house lawyers are more […] aware of how automation can cut down on costs and increase productivity,’ she says. ‘That means they can take care of things in-house as opposed to sending it to expensive law firms.’

Yet while law firms have been experimenting with and investing in such tools for several years, the picture on the in-house side of the equation appears to be quite different. An in-house technology lawyer at a UK-based wealth management business, who wished to remain anonymous, says that while for law firms there’s a clear business case for automating processes, for in-house legal functions the investment required to get there can make automation a far harder sell to those running the business. ‘From our perspective, we’re considered a back-office support function,’ says the lawyer. ‘Because [the legal team is] not revenue generating […] we struggle to secure resources. If you look at all other areas of our business, technology is being harnessed in all different ways and we’re seeing much more automation driving efficiencies, but to allocate resources and time and knowledge for the automation of the legal function it’s quite difficult to get the buy-in and make the business case.’

Fredrik Hallin, Head of Legal at Stannah Management Services, agrees, and says that sometimes that’s because incorporating the technology required to automate legal processes is seen as unnecessarily disruptive to the business as a whole. ‘To incorporate some document-management systems in a corporate environment would require a significant investment because you would have to reconfigure the entire IT system and businesses can’t perceive the benefit of that,’ he says.

For Hilber, that creates an opportunity for external counsel to become even more valuable to their clients. Indeed, rather than revolutionising in-house legal departments in the sense of changing either their make-up or the way they work, the automation of certain processes has simply enhanced what firms can offer in-house counsel without causing costs to soar. ‘A lot of our clients rely on us for these tools because they don’t have them themselves,’ he says. ‘Often it is complicated for them to have the software installed in their systems but we can let them have it on a cloud-based basis. We offer this as a free service. It’s nice and convenient.’

‘What we see is that in-house counsel are really happy if we can provide them with solutions on the basis of AI and automation,’ adds Hilber. He says this has enhanced relationships with many of the firm’s clients, without taking anything away from the service they provide. ‘Five or 10 years ago we would have charged for an NDA or simple documents that we do for clients. That’s not really what we do though so the relationship has become more focused on things that really matter,’ he explains.

Generative AI changes the game

That doesn’t mean that revolution is not in the air, though. While there’s been much talk about AI in recent years, when it comes to the legal sector the conversation has largely referred to machine-learning style document review. Generative AI – artificial intelligence such as ChatGPT that can produce rather than simply review documents – puts a completely different complexion on things.

For a report published in August this year, Thomson Reuters, which develops and supplies software solutions, surveyed over 1,200 people working in professional services – tax, accounting, risk and compliance professionals as well as lawyers – across the UK and the Americas. It found that the majority believe their professional lives are on the cusp of seismic change, with AI, automation and other technology expected to have by far the biggest impact on their businesses in the next five years.

In the foreword to this document, the Future of Professionals Report, Thomson Reuters Chief Executive Steve Hasker wrote that much of that is going to be driven by generative AI. ‘Many of the innovations we use today have crept up on us, evolved, and gradually changed the way we work and play,’ he wrote. ‘But not generative AI — this is different. It has captured our collective imagination and changed everything in a very short amount of time […] The easy interface of recent generative AI tools has made us all suddenly aware of its immense capabilities.’

Hasker believes it’s critical that we acknowledge AI’s shortcomings. Yet, he argues, if AI tools are built and harnessed correctly, they have the potential to overcome some of businesses’ biggest pain points. These he lists as overcoming shortfalls in the labour markets; increasing job satisfaction and mental health; automating the most unpopular and menial of tasks; and helping professionals navigate the increasingly complex regulatory and compliance landscape – without adding cost or time. This list, in Hasker’s view, ‘is just the beginning’.

Hallin agrees. He says that so far in the legal sector, automation has only been used where it makes absolute sense to use it – for example, for low-stakes document production, rather than the creation of high-level contracts. However, next-generation AI offers the kind of opportunities that haven’t yet been seen in the profession. ‘We talk about why lawyers are slow to adopt new technologies and my general sense is that lawyers are quite happy with what they have but I think it’s about to change dramatically with AI,’ he says. ‘We’ve had a slow burn with contract automation and project management tools but we are up for a dramatic change.’ Business models, he adds, will look very different, very soon and the billable hour charging model will be affected in the context of further AI-driven automation of legal drafting, review and research.

For Hamblin, it’s the shift from machine learning to AI models that’s going to prove the real game-changer for the legal profession. Trying to teach computer systems to work like humans didn’t have the anticipated impact on legal processes, but now that machines can use algorithms to replicate human intelligence he believes that AI can be much more use to clients and law firms alike. ‘AI that used machine learning didn’t work,’ Hamblin says. ‘For due diligence we spent a lot of time trying to train it but it was a waste of time. The great game-changer is model AI. If you’re using the right underpinning model then established players will be able to put something on the market that will be very attractive. It will be totally bespoke and will use a database of precedents to understand how a lawyer thinks a bit more.’

Hamblin doesn’t believe AI will be able to generate the kind of documents lawyers currently spend time writing – there’s too much nuance in what humans are able to learn and process for that ever to be fully replicated by AI. However, he does think that AI will be able to standardise what humans produce, something that’s likely to enhance the relationship between in-house lawyers and their external advisers. ‘Some lawyers are more verbose than others and an in-house lawyer might find it tedious going through lots of sentences and paragraphs from their law firm,’ he says. ‘If they said “can you be more succinct in your writing style” that lawyer would be able to say I’ll write it in a way I feel comfortable with but I’ll use AI to turn it into something the client will read […] they could write it out and then say summarise it in bullet-point format.’

Rethinking roles

AI will put new demands on external counsel, says Hamblin, but it’ll give them less excuse for providing work that’s not top-quality. It’ll better allow them to tailor their advice to requirements too. ‘It will enable all of us to be rocket scientists – we won’t have knowledge of every area of law but we’ll be able to access it more easily,’ believes Hamblin.

Hallin explains that this will necessarily change the relationship between in-house lawyers and their external advisers, though it’s not yet clear how that will manifest itself. The impact it has on private practice firms, which have already had to rethink the role of associates as a result of the opportunities machine learning has presented, could be wide-reaching. However, Hallin doesn’t believe it will lead to in-house roles being replaced. ‘There will be certain tasks where it will be more efficient to rely on a machine,’ he says. ‘I’m thinking the role of the associate will look different because the basic tasks that they do will be quite easy in the next couple of years to replace.’

But just as Hamblin cautions on just how far-reaching generative AI is likely to be, Hilber says there’s always going to be a need for human intervention, even with the step change that’s likely to come. ‘So far AI-driven solutions have not delivered the desired results or the expected results but the solutions will get better and clients will expect their firms to use it to be more effective,’ he says. ‘The issue is reliability. We can’t really see how reliable the results provided by AI will be at the moment. If lawyers use AI without checking the results, it could be disastrous.’

At the moment, AI is being used to provide summaries, to create PowerPoint presentations and to have a starting point for work – without replacing lawyer input, essentially. Hilber is sure that sooner or later AI will become more effective, but how widely it’ll be used to replace lawyers depends upon which matters it’s being used for and whether there will be insurance cover for mistakes made by AI. ‘An AI provider will never be able to take on the liability and if the law firm takes on the liability it will need to take responsibility to double check [the work produced by AI],’ he explains. ‘Insurance for AI solutions could be the really interesting game-changer.’