By Ben Williams, Pinsent Masons Vario
The way we look at adopting generative AI tools needs to change.
Cast your mind back just five years, the process of introducing a new legal technology solution would start with multiple teams sitting down together and pinpointing exactly what their requirements were before carrying out an exhaustive market review to find the best possible solution. That approach is no longer fit for purpose.
Pilots in key areas more effective
Across the in-house legal teams that we work with, we’re finding that it’s now much more effective to pick a few areas where you feel there’s real potential for improvement and start experimenting; ideally with a tool that you either already have available or can procure very quickly and without vast expenditure.
After all, one of the benefits of generative AI is the tools being incredibly quick to introduce and accessible for the mainstream workforce.
Then, once you have these in place, we recommend focussing on delivering practical, high impact wins across your focus areas. This will hopefully improve familiarity, confidence and buy-in from the legal teams that you work with and highlight where further gains may be made.
Of course, this is an iterative process and I’m constantly finding new areas where I feel generative AI could make a real difference for clients.
Speed of market change outpacing traditional project timelines
Of course, this approach does mean that you might spend time and money on solutions which turn out not to work as you’d hoped.
My response to that challenge would be to say that’s fine, and in the nature of experimentation. The AI market is moving so quickly that by the time you have completed a more traditional approach, you face the same risks anyway; the technology will have changed and made your work obsolete.
Having experimentation and education as your primary focus will leave your team better equipped to spot the right opportunities and move fast themselves, which will in turn open up more significant opportunities. In time, the benefits that you realise will outweigh the cost of early experimentation.
Being hyper-focussed on what your problems are and finding pragmatic, if imperfect, solutions today will stand you in good stead in the longer run.
Legal-specific AI solutions better for serious legal work
One of the questions we’re often asked by in-house legal teams is whether they need to invest in a legal-specific AI tool or should introduce a general-purpose model.
At the moment, you can do more that you’d expect with general-purpose tools like Copilot, especially logistical and organisational tasks, but legal-specific AI tools are still outperforming general-purpose solutions when it comes to serious legal work.
A lot of ink has already been spilled over whether legal-specific AI tools will eventually be made obsolete by the likes of Anthropic and Microsoft making serious plays in the space. There’s no doubt that they’re beguilingly impressive, and will have an impact, but it’s still too early to say exactly what this will be. There is a long way for this race to run and the winners are unclear, but that cannot be an excuse for inaction.
On choosing solutions, thinking beyond the product to whether a vendor company has good leadership, solid funding, real clients and a clear plan for the future can help you select a tool with real longevity.
Reminding users of potential gains crucial
When in-house legal teams tell us that they’re struggling with low solution adoption, we find that the issue is rarely the technology itself.
Adoption is about making a start, building familiarity, debunking myths and developing trust. The sheer pace of change that we’re seeing does create anxiety, as does a perceived lack of transparency and accountability; lawyers are trained to be sceptical after all!
Many evangelists also exacerbate these issues when their claims turn out to be overly optimistic. We have to remind people that the potential gains are real and significant; waiting to see how the market develops while your competitors move forward isn’t an option.
Learning and real impact more important than perfection
Moving quickly can mean buying a tool, or indeed several tools, setting up a working group, or meeting lots of vendors.
We’re still really early on this journey. For me, moving quickly means not waiting and not aspiring to perfection. Experiment, try things. Right now, learning is more important than measurable gains and real impact is more important than being right at the cutting edge.
Ben Williams is head of consulting and technology at Pinsent Masons Vario. He has previously held senior management roles at KPMG, Elevate, and Allen & Overy Peerpoint.
Unlike many other publications, we don’t charge for guest posts or any written editorial content – if you have something interesting to say, message us. [email protected]