In-house lawyers’ AI usage revealed - but are you a truck or a hybrid?

The results from our annual survey of Juro community members are out (read them in full at the State of In-house 2024: legal teams in the age of AI).?

The picture has changed significantly in 12 months. This time last year, from a cohort of 100+ in-house lawyers surveyed, 44 per cent told us they would not be using generative AI tools at all. Fast-forward to today, and around the same proportion told us they’re using it weekly or daily:

We’ve talked here before about legal’s suitability for disruption by generative AI, with Goldman Sachs’ provocative stat (that 44 per cent of legal tasks lend themselves to full automation) having a strong weight behind it.

Anyone with experience of legal work knows that much of it involves parsing large amounts of text to draw conclusions and provide concise advice based on what’s been read. LLMs are superb at this, even if not perfect.

It’s no surprise, then, that tech-forward lawyers adopting the latest AI are turning its cannons towards those time-consuming tasks that fit well - contracts, drafting, proofreading, summarising:

Against that backdrop, vendors still have work to do to prove that AI is safe and trustworthy:

… and that the ROI is sufficient, and (crucially) easy enough to realise, for them to stop dabbling and start embedding. For platforms like Juro, we have to make sure the tent is big enough to welcome every type of AI user (and non-user). Thankfully we have no shortage of market feedback and customer data to make good on that promise.

We’ll be talking through these key findings, as well as others relating to CLM adoption, collaboration with the business, and even stress and burnout, in a webinar in a couple of weeks - click here or on the image below to join us ??

19 per cent using gen AI every day, and 25 per cent every week, is eye-opening. But what do we mean by ‘using’? After all, there’s a difference between eating vegetables and going vegan.

From our sessions at Scaleup GC, we know that in-house lawyers are finding applications of AI for a whole range of purposes beyond contracts - data visualisation, writing training materials, reading through codes of conduct, and so on.?

After all, there’s a difference between eating vegetables and going vegan.

But it’s worth remembering that in the context of an epochal technology, we’re still in the early days of AI adoption - perhaps even earlier in legal than in other sectors too. A bit like our friends at Point Nine , who use animals to describe customer size, we are building our own taxonomy for AI users.?

• people with no interest in ever using it (let’s call them trucks);
• people who’ve dabbled from time to time (petrol cars);
• people who use it regularly, perhaps weekly or even daily (hybrids); and
• people who are all-in on AI, embedding it in their workflows wherever possible (electric vehicles)

Many lawyers - perhaps even most, in our community - have given ChatGPT or Claude a whirl for a first pass at a clause rewrite, or for a first draft of a particular email when they’re pushed for time.

But how many are actually embedding it into their workflows? With incoming legal queries passing through AI triage by design? Or sales teams being served negotiation fallbacks by AI guardrails? Or AI chatbots handling routine queries, self-serving from the company’s knowledge base in Notion or Confluence or wherever else you keep your in-house knowledge?

Despite all the noise, usage data of Juro’s AI Assistant, as well as our survey data, tells us that a healthy chunk of legal’s AI users are at the petrol car stage and not yet even ready to try the hybrid. The true electric vehicles remain few and far between. For now, at least.?

To stretch the analogy even further: various governments have mandated that the production of petrol cars must cease in the medium term. I strongly feel that AI will be as foundational to legal tasks as electric vehicles will be to transport. The question isn’t if, but when.

What do you think?