
Talk
Your agent answered 1,000 RFQs overnight. How many were right?
Everyone in this room expects AI agents to take over parts of B2B sales. Procurement agents already compare suppliers and request quotes, and Gartner expects 90% of B2B purchasing to be agent-intermediated by 2028. Sellers are answering with agents of their own.
The hard question comes after the demo. How do you know your agent quoted the right price, picked the right part and promised a delivery date you can keep? Across borders and languages, mistakes are even harder to spot.
Based on current projects with European manufacturers, this talk covers three things: what a B2B sales agent needs to work, from context to live access to prices and stock to clear rules; how to test it against real historical cases before go-live and monitor it in production; and who owns it day to day. Attendees leave with a five-question checklist for their own agent readiness.
One example from practice is replaying several hundred past quote requests against the agent and comparing its answers line by line with what the sales team actually sent.
Key takeaways:
• Which data, system access and rules a B2B sales agent needs before it answers its first customer
• How to build a test set from your own historical RFQs and orders, and how to read the results
• Where to set the point at which a human takes over, and who in the business owns that decision
About
Matthias Dietrich is the founder and Co-CEO of foobar Agency in Munich. foobar Agency helps enterprises become agentic by advising on, planning and building AI agents for sales and for steering the business on data platforms. A former developer, Matthias has spent more than 20 years on digital sales platforms for retailers and manufacturers in Europe.