Employee grievances, customer complaints and GDPR requests are now routinely drafted using AI tools, which has placed a significant burden on in-house legal teams.

The report, From AI Adoption to AI Advantage, is based on in-depth interviews with senior in-house lawyers at companies in Ireland, including eir, Flipdish, Mastercard, Gilead Sciences, Uniphar and Wayflyer.

The research was conducted by Dentons’ Dublin-based technology, media and telecommunications partners Carlo Salizzo and David Kirton.

Most of the companies interviewed said that the issue is not reading this AI-generated correspondence but rather working out which parts of it matter. Some in-house teams have turned to AI to deal with the sheer volume of AI-generated submissions.

Almost every interviewee said AI was being used for narrow, well-scoped repetitive tasks, rather than being given free rein to take on legal work. Many companies emphasised the importance of centralising legal data. A number use a wide variety of AI tools to work with that data.

Mr Salizzo said: “This research shows that in-house legal departments are not replacing employees with AI, but they are finding ways to use AI to help staff be more productive, including using AI tools to help cope with a raft of AI-generated submissions.”

Mr Kirton said: “The days when governance was the exclusive territory of legal departments are over. This research shows that most companies have found that the advent of AI has meant that good governance requires the expertise of multi-disciplinary teams.”

The report also identified significant changes in the relationship between external counsel and in-house lawyers, with the latter exploring alternatives to the billable hour. Some argue that it may die out, while others see alternative models as a way of passing efficiency gains back to clients.

More broadly, the nature of the client-lawyer relationship is also changing. Clients are now more rigorously prepared, leading to more intellectual sparring with outside counsel over the difficult technical challenges that AI can’t solve.