AI Under the Hood: What Every In-House Lawyer Needs to Know About How LLMs Work

Cecilia Ziniti

You’ve used AI for legal work. Sometimes the results feel like magic. Other times, they fall apart mid-response, hallucinate a case that doesn’t exist, or lose track of your instructions entirely. The difference between those outcomes usually comes down to how the technology processes information, and most lawyers have never been shown what’s actually happening behind the curtain.
Join us on Wednesday, April 9, 2026 at 12:00 PM ET / 9:00 AM PT as GC AI’s Solutions Attorney Brittany Pfister and a GC AI AI engineer break down how large language models actually work, where they go wrong, and what in-house lawyers can do to get consistently better results. No computer science background required. We’ll explain the mechanics in plain English, then turn that understanding into techniques you can apply immediately.
Key Takeaways:
How LLMs process your prompts: What’s really happening when you send a question to an AI tool, and why small changes in how you ask can produce dramatically different outputs
Context rot and how to prevent it: Why AI loses the thread on longer tasks, and practical methods to keep your results accurate from start to finish
Reducing hallucinations: Concrete techniques to catch, prevent, and minimize fabricated outputs in your legal work
Agentic AI, demystified: What it actually means, how it works, and why it matters for the future of in-house legal workflows
Who Should Attend
This program is designed for in-house counsel who are already using AI in their practice and want to move beyond trial-and-error. Whether you’re drafting contracts, reviewing vendor agreements, or summarizing board materials with AI, you’ll leave with a deeper understanding of the technology and advanced techniques to get more reliable results every time.
There is NO COST to attend this program!
This program is FREE, thanks to our sponsor, GC AI. Can’t make it to the live program? The recording will be available in our Free Video Library.



