Agentic AI
Agentic AI is moving from prototype to production faster than enterprise data platforms can keep up. Agents reason, plan, and act across many data sources at once, but the data they touch was rarely designed for them.
This workshop is about that gap. How do we make enterprise data usable, auditable, and trustworthy for autonomous agents?
What we will dig into
Reliable access to structured data. Text-to-SQL, hybrid retrieval, and what to do when an agent writes a valid query with the wrong answer.
Auditability and provenance. Tracing what an agent read, synthesised, and changed, so decisions can be reconstructed.
Data quality for agents. Moving from static rules to continuous, context-aware quality that agents can interrogate before acting.
Who should come
Anyone working at the intersection of enterprise data and agentic AI. Engineers, architects, researchers, governance leads, product people, students. Academia and industry equally welcome.
Format
Technical workshop, not a vendor showcase. Speakers from research and from production, side by side. Expect deep talks, honest experience reports, and time for real discussion.
Speakers:
🎤 Petar Tsankov, CEO and Co-Founder at LatticeFlow AI: The Governance Gap in Agentic AI. How to Close It.
🎤 Tobias Häberlein, Head of Department Computer Science, Fernfachhochschule Schweiz, Context Is the Bottleneck: Context Engineering as a Response to the Enterprise Data Gap.
🎤 Peter W. J. Staar, Manager of "AI for Knowledge" group in IBM Research Zurich: From Unstructured Documents to AI-Ready Knowledge: Docling, DocLang, and Enterprise Document Intelligence.
Hosted by DBTA, https://swissinformatics.org/dbta.php?Lang=de, in partnership with SBB.