How Thomson Reuters Built AI Agents That Think Like Lawyers

In this episode of the Neuron AI Explained podcast, Corey Knowles and Grant Harvey interview Joel Hron, CTO at Thomson Reuters, about Deep Research, an AI agent platform designed for legal strategy. Joel explains how Deep Research differs from general AI tools like ChatGPT by being specifically tuned for legal research, incorporating legal reasoning, and utilizing Thomson Reuters' extensive legal content and tools like Keysight. He highlights how Deep Research can significantly reduce research time, uncover new arguments, and augment the capabilities of attorneys. The discussion also covers the importance of human experts in the loop, the explainability of AI in legal contexts, and strategies for mitigating hallucinations. Joel shares insights into Thomson Reuters' model-agnostic approach, their excitement about open-source models, and the challenges of making cutting-edge AI enterprise-ready for legal professionals. He also touches on future applications of AI agents in tax and the importance of adaptability in the rapidly evolving field of AI.