The Treasury Update Podcast Intro to Kyriba's TAI
Sep 15, 2025
Tom Gavaghan, a product leader at Kyriba specializing in treasury solutions, discusses the groundbreaking concept of Trusted AI (TAI). He breaks down the differences between agentic AI, generative AI, and machine learning. Tom emphasizes the importance of treating AI like a new hire, advocating for cautious implementation and continuous training. He underscores the need for high-quality data to ensure optimal performance and explains Kyriba's focus on security and compliance. Listen to his insights on how TAI could redefine treasury operations.
AI Snips
Chapters
Books
Transcript
Episode notes
AI Evolution: From Models To Agents
- Agentic AI uses generative AI but operates autonomously to run processes and make decisions.
- Tom Gavaghan frames AI evolution as machine learning → generative AI → agentic AI, with agents working together to achieve goals.
Agentic AI Will Become Mission-Critical
- Agentic AI will become mission-critical in treasury over time as adoption and trust grow.
- Early stages require pragmatic rollout and human oversight while capabilities mature.
Hire AI Like You Hire People
- Treat AI like hiring an employee: start at appropriate job levels and monitor performance closely.
- Incrementally evolve and coach AI agents rather than handing over critical duties day one.
