Harnessing Unstructured Data for Agentic AI Context
Unstructured data holds over 80% of what your enterprise actually owns, and it’s still AI’s #1 blind spot. Kevin Petrie, VP of Research at BARC, and Kyle DuPont, CEO of DataXray, discuss how to govern these files and make them AI-ready.
See where your unstructured data stands on AI readiness
Get the recording
Four decisions behind an AI program that scales
Close the board-level gap
79% of organizations say they can extract value from unstructured data safely, but only 29% can actually find the right files for a given use case.
Knowing which side of that gap you sit on lets you commit to an AI roadmap you can defend, rather than discovering the shortfall halfway through a board-sponsored rollout.
Stop pilots stalling at scale
The pilot that impressed everyone on 200 curated files falls apart across 20,000 SharePoint sites, because top-K retrieval rarely returns the right chunk from a billion.
Fixing that with a context layer rather than more infrastructure is the difference between a program that scales and another eight-figure platform bill.
Say yes to the business
74% still govern unstructured data by hand, which caps how many teams IT can support and makes the function look like the bottleneck.
Automating classification and access at file level turns that around: business units get approved in days, and the exposure you carry is decisions made on the wrong documents, not just the audit finding.
Sequence spend correctly
Only 17% are vectorizing, and the leaders are still on classification and validation first.
That ordering is worth defending in a budget review, because an agent that knows what a file is will do the retrieval work itself, and the cheaper foundation ends up carrying more use cases than the expensive one.
Who you will hear from
Kevin leads the data management practice at BARC, writing on AI, data integration, and data governance. For 30 years he has worked out what technology means for practitioners, as an industry analyst, instructor, marketer, services leader, and tech journalist. He built a data analytics services team for EMC Pivotal across the Americas and EMEA, and ran field training at data integration provider Attunity, now part of Qlik. A frequent speaker and co‑author of two books on data management, Kevin’s main interest is teaching data and AI leaders how to capitalize on the value of their data.
Kyle co‑founded DataXray after watching enterprise after enterprise hit the same wall on AI: the files holding their most valuable information, the documents, emails, and contracts AI needs to work, were the files no tool could open, read, or classify. He has worked with major financial institutions, government agencies, and defense organizations to bring that data into view and make it ready for AI, compliance, and security workflows. DataXray reads and classifies files across the enterprise estate, feeding the results into the AI, compliance, and security tools already in place, without moving the data.
Unstructured data in my view holds the key to creating a differentiated version of agentic AI… Trustworthy facts come from tables in a lot of cases, but the nuance, the true context, comes from unstructured data.
Agentic AI runs on context, and your context is in your files
Watch the session for what it takes to get to it.