Many organisations are building frameworks of interoperable data through data products, data contracts, and data marketplaces to make their data findable, accessible, interoperable, and reusable (FAIR). With the adoption of standards for data products and contracts – and their publication in data marketplaces – these assets become not only discoverable by humans, but also programmatically accessible. This creates a scalable path toward AI‑ready data and automation.

Data products provide access to high‑quality data along with essential context such as definitions, ownership, governance, and usage rules. However, this context only describes the data product itself. It does not capture the wider organisational context: how the organisation works, how decisions are made, what assumptions are baked into operations, and how value is created.

So how do we capture this institutional knowledge – our organisational memory – in a way that enriches automation and enables AI systems to act in alignment with company principles?

In this session, we will explore the key capabilities enabled by data products and invite participants to discuss (or debate!) how organisational knowledge can be modelled, shared, and applied to agentic AI systems. Attendees will work hands‑on to map data products, identify missing context layers, and design lightweight mechanisms for capturing and operationalising institutional knowledge. The session blends theory, templates, and peer coaching to help participants move from abstract data‑culture aspirations to actionable design practices.

Moderator: Noemi Moreno Fabelo, Head of Data Strategy & Architecture | Cambridge University Press & Assessment