Understanding True AI Agents in Legal Tech

In the latest episode of Law Punx, Jake Jones, co-founder of Flank, discusses the widespread misunderstanding of AI agents in the legal industry. The 8-minute episode delves into what constitutes a true AI agent, how they function, and why many legal professionals misclassify automation tools as agents.

Key Insights:

• Autonomy vs. Automation: True agentic AI can autonomously make decisions and take actions, unlike linear orchestration tools that simply execute predefined steps.
• Scope of Work: Agents handle entire workflows, not just individual tasks, and can adapt plans based on real-time feedback.
• Human Oversight: While agents operate independently, human supervision remains crucial.

Jake emphasizes that many tools marketed as "agentic AI" are merely decision-tree automations, lacking the ability to self-reflect or adjust plans dynamically. He provides examples, such as a hotel booking agent that negotiates and adjusts based on user preferences, contrasting it with static document review tools.

Why Lawyers Misunderstand Agents:

• Paradigm Shift: Legal professionals often view AI as assistants rather than autonomous systems capable of handling entire workflows.
• Domain Expertise: Many legal tech founders build tools based on their experiences as lawyers, limiting their ability to envision fully autonomous solutions.

Key Takeaways:

1. Agentic AI autonomously makes decisions and takes actions.
2. Misclassification of tools as agentic AI is rampant.
3. Human oversight is essential even in autonomous systems.
4. Future of Legal Tech lies in fully autonomous AI systems interfacing directly with non-experts.

For more, watch the episode here or listen on Spotify or Apple Podcasts.

AL TV Productions, 2025.