Most enterprises trying to put AI into production hit the same wall: the data is not ready. The instinct is to point a frontier model at the documents and hope for the best. It works once, then falls apart at scale.
In this webinar recording, Alan Jacobson, Founder, Advisor & ex-CDAO at Alteryx, joins Flexor founder Or Zabludowski to work through what actually has to happen to unstructured data before an agent can use it reliably, and what it costs when you skip those steps.
Would rather read than watch? Read the full transcript.
Webinar recap: from unstructured data to AI context
- Why preparing data for AI mirrors how a person prepares for an exam, and why the answer changes completely between one page, one book, and an entire library
- Shifting work left: moving indexing, caching and context building off the frontier model and into a dedicated engine
- A live demo on a single 112-page report, dense with tables, cross-references and charts, where contextual chunking returns the same answer for one ninth of the tokens
- A second demo across 100 SEC filings, comparing an agent running Flexor’s Model Data Protocol against RAG on the identical query
- How a Domain Intelligence Hub keeps terminology consistent when a large agent spins up smaller subagents
- What an open architecture looks like in practice, with Iceberg tables and a query engine as the gateway for agents
Audience questions answered
Why a PDF parser is not enough on its own. What you can do with unstructured data today if you are not running agents yet. How organizations should prepare, on the team side as much as the technology side.
Speakers
Alan Jacobson, Founder, Advisor & ex-CDAO at Alteryx
Or Zabludowski, founder at Flexor
See what this looks like on your own data
Book a demo and we’ll show you how Flexor can add value to your unstructured data use cases.
