This is obviously critical for us – clients come to our firm for quality of advice and the assurance that we’ll deliver. The first part of the answer is a human one – we remain individually responsible for the work we send to our clients and we have a comprehensive training programme that supports that. We’re also designing our workflows and increasingly our agentic systems in such a way that they require close cognitive engagement with the output. Secondly, quality output depends on giving the right context to the models, so having access to leading tools grounded in curated datasets has been critical. Finally, the tools themselves can contribute to a better outcome – automated checks of case references, agent flows that verify the work from different angles and guardrails around the sources used. Ultimately though, it all comes back to hiring and developing people at the top of their game for whom the technology is a tool and not a substitute.