Just 6% of firms have deployed AI at scale
Many organisations may feel they’re being left behind with generative artificial intelligence (GenAI), but the good news is that it’s not too late yet for legal professionals to get involved and make it work for them, delegates at the recent Law Society AI in Legal Practice Summit were told.
“Companies are just at the beginning in terms of creating the structures and processes that lead to using GenAI effectively – there are virtually none out there yet that are doing this really, really well,” said futurist and author David Kerrigan in his keynote address (26 March).
“I think it’s important to realise that, as AI is moving very fast, within that people are moving at a different pace. So, you’ll have some firms that take on AI – they’ll buy into it, embrace it and invest in it. And there are others that will take a more cautious, slower approach. But more granular than that, within organisations, people are moving at different paces. It’s a very individual thing.
“So, don’t be discouraged if you’re feeling overwhelmed and thinking ‘I can’t do this, it’s too hard, I don’t even know where to start.’ There’s still time to do it and, if you redesign your workflows and put the right governance in place, it can work,” he added.
In her presentation, Donna O’Leary, founder and director of DOL Consultancy, noted that 98% of organisations had started with AI, with most still piloting tools, according to PwC Ireland’s GenAI Business Leaders Survey 2025.
However, only 6% have deployed AI at scale and 79% lack formal adoption structures.
“There’s this consensus in the market of waiting to see what other firms adopt, especially the big players. In parallel to that, you absolutely need to just be doing basic AI literacy training that is tech-agnostic and covers use cases, limitations, professional obligations, and the legal and regulatory environment,” said O’Leary.
Importance of AI literacy training
O’Leary shared two case studies to give delegates pointers on things to watch out for if they’re at the stage of potentially piloting a GenAI tool for their firm.
The key point illustrated was that AI literacy training, before any tool is even selected, can have an impact on adoption rates.
The first law firm, employing 180 staff, had adopted a legal-specific AI tool across the firm after a one-month pilot.
The AI provider gave staff standard on-boarding training on the tool but did not teach them the fundamentals of AI.
“On paper, the approach of the first law firm seems reasonable. However, adoption rates were lower than expected, at 40% on average, and not all functionalities of the tool were being used. Employees struggled to find applications for the tool in certain parts of their work. That was the reason I got pulled in,” O’Leary explained.
“Client feedback wasn’t great. For example, clients were receiving documents and noticed issues with them – not hallucinations per se, but misunderstandings and misinterpretations of the information being presented to them.
“Basically, the staff weren’t verifying the outputs because they weren’t trained on the uses and limitations of these tools, and what to watch out for.”
On top of this, the firm subsequently realised that the tool wasn’t compatible with its case-management system, which emerged only after the fact.
Towards successful implementation
The second case study was of a law firm with 150 staff that engaged O’Leary from the outset. She provided them with tech-agnostic AI literacy training, covering everything they’d need to know before bringing a pilot to the firm – including the functionalities to expect from any legal-specific tool.
The firm tried four different tools but quickly narrowed them down because it knew the right questions to ask from the outset, having adopted a vendor checklist for key requirements – including compatibility with the case-management system.
After a one-month pilot, the tool was rolled out across the firm.
From the pilot, AI champions were identified in each practice area or department – a step O’Leary indicated was crucial to successful implementation because it helped embed the tool across teams and maintain momentum after the initial roll-out.
Vendor training
“The staff did receive on-boarding training from the vendor, but they continue to receive training on the tool, which is important. One and done is not enough. Firms need to put continuous training on the table and be upfront with vendors to have it included in their package, even at an extra cost.”
The result of following these steps was that adoption rates were significantly higher because people were enabled and educated, O’Leary said.
“They knew what questions to ask before the pilot began and were keen to get stuck into using the functionalities, such as playbooks. AI champions are moving up to more advanced features and feeding that back to junior and other staff members who aren’t as far ahead. And they continue to report to senior management, because AI adoption is a cycle that never stops.”