Testing Agentic AI systems in the real world
Agentic AI holds enormous promise. But first, it must prove it is reliable enough for real-world deployments.
From exciting research and impressive demos, agentic AI might well be making its way into many real-world products and services soon. But if you are thinking of deploying it, you are probably also asking yourself:
👉 Can it perform consistently—with correctness, safety, and security—at scale?
Join us for the next AI Verify Foundation Community meetup.
🔍 What’s on the agenda:
What is agentic AI? How do we define it beyond the buzzwords? (Spoiler alert: there isn’t a standard one!)
How is it being really used in organisations? Why is there such a gap between personal GenAI adoption and enterprise use?
What are the real-world risks? From compliance to misuse, what do technical and business teams need to prepare for?
(How) Does Testing looking different for Agentic AI systems? Or is it just a question of extending and deepening the scope of “traditional” LLM app evals?
We will bring these to life through a mix of
Short sharing sessions from 3-4 AIVF members who have started implementing and testing agentic systems
The outcome of the recent Agentic Reliability Evaluation Summit organised by our partners, ML Commons, in Oxford
Results of an informal survey we ran with our extended community worldwide – on the real extent of adoption and the way in which risks are perceived and managed
Lots of informal conversations!
👥 Who should attend:
Non-technical business and product practitioners looking to understand the real state of Agentic AI adoption in Enterprises, and how pioneers are approaching the risks
Technologists seeking to understand what it might take for them to win the confidence of their stakeholders as they seek to roll out Agentic AI in their organisations
🎯 Let’s move beyond the hype and talk about how to build agentic AI that’s not just powerful—but reliable.