AI in the legal sector is moving beyond simple automation into something far more transformative. In this podcast, Legal IT Insider’s editor Caroline Hill and Harvey legal innovation partner Tara Waters explore the rise of AI agents – tools that don’t just execute instructions but can plan, reason, and complete multi-step legal tasks with a degree of autonomy. As Waters explains in this relaxed conversation and exchange of ideas, agents can take an outcome like producing a due diligence report and determine for themselves how to get there, marking a significant shift from traditional software.
These capabilities are already being put into practice. On platforms like Harvey, hundreds of agent use cases are emerging, from straightforward tasks like checking documents against due diligence lists to more complex workflows such as generating full sets of legal documents by gathering context and drafting outputs. While still in early access, these tools show how legal work can be streamlined, and in some cases, fundamentally redesigned.
This new level of autonomy brings important challenges. Lawyers are rightly cautious about ceding control, especially when agents can make decisions or take actions independently. This has made governance a central concern, with firms needing to carefully manage data access, system integrations, and risk. Waters and Hill discussed how to approach agentic AI adoption: the recommended approach, much like with all new technology, is to start small, testing agents on low-risk tasks and gradually expanding their use as confidence grows.
These are exciting and challenging times in equal measure, and Waters discussed how AI agents do not remove the need for lawyers, they increase it in different ways. Human oversight remains essential, with lawyers responsible for designing, testing, and verifying agent outputs. Legal judgment, rather than being replaced, becomes even more important in assessing whether AI-generated work meets the required standard.
Looking ahead, the conversation points to a near future where legal teams operate as hybrids of lawyers and AI agents working in parallel. What do firms need to think about right now to be ready to adapt?
Listen to the full podcast on Spotify and YouTube below:
