Procurement AI Foundations: Intake, Data, Rules

Most procurement teams already have a Source-to-Pay platform. Yet requests still arrive by email, spend sits across ERPs that do not agree, and much of the real work happens in spreadsheets around the system rather than inside it. Agents can only take on that work once the foundations underneath them are in place.

This session looks at what has to be true before an agent can carry a request from intake through sourcing, contract and payment: a front door people actually use, spend and supplier data that can be trusted, and clear rules on what an agent may decide on its own.

Hosted with CIPS and moderated by James Moore, Editor of CIPS Download, the session brings together two speakers. Raoul Gröning, Lecturer in Procurement at Kalaidos University, explains why AI adoption in procurement lags and what every agent needs before it can be trusted. Pranshu Sahay, Solutions Consultant at Zycus, follows one request end to end on the live platform.

What you’ll learn

Why teams work around a platform they have already paid for, and why intake is usually the first thing to fix
What a simple front door for requests looks like, and why it decides whether people use the system at all
What has to be cleaned up in spend and supplier data before agents can act, and what can stay exactly where it is
The rules every agent needs: a named owner, clear limits, approvals, an audit trail and a way to stop it
Why you can hand an agent the work, but not the accountability
One request followed live from intake through sourcing, contract and payment, with the agent handling the steps it has earned

The session ends with a live demo and questions from the audience.