As “deploy fast” meets enterprise reality, data governance has quickly emerged as the bottleneck between AI ambition and outcomes.

The disconnect is measurable. In their Agentic AI Study, Qlik Technologies Inc. found that while 97% of enterprises have committed budget to agentic AI, only 18% have fully deployed it — with data quality, integration and governance cited as the leading blockers. The AI-first workday may be the destination, but trusted, governed data is the only sustainable path there, according to Sam Pierson (pictured), chief technology officer of Qlik.

“I think the biggest thing that we see as these things go into pilot mode, or whether they’re trying to move into production, [that] the biggest constraint that these organizations have is the data,” Pierson told theCUBE. “You think about all of the systems they have in-house — whether it’s a small company or a large enterprise — there’s hundreds or thousands of systems, databases that have been around for a long time that maybe haven’t been modernized, and it’s like, okay, how do we get that data out of those systems [and] into our AI reference architecture?”

Pierson spoke with theCUBE’s Rebecca Knight and Rob Strechay at Qlik Connect 2026, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed data governance strategies for scaling AI, the role of Model Context Protocol in enterprise deployments and how a semantic layer unlocks trusted intelligence at scale. (* Disclosure below.)

Expanding access to data and AI tools across an organization demands a deliberate governance foundation — one that doesn’t just live in policy documents, but is embedded in the architecture itself. When organizations begin working with MCP initiatives and similar standards to move beyond siloed experiments, data governance quickly becomes a security conversation, Pierson noted.

“The first conversation you have is, ‘How can you help me with AI?’ We checked that box. Now the next conversation is with the security team and they want to know, ‘How are you securing the data so that we don’t have a data leakage issue?'” he said. “Our AI really inherits its security model from the core platform.”

That security model extends into lineage and observability, with Qlik’s semantic layer providing metadata context across data products and upstream impact analysis. Rather than chasing the latest model or architecture pattern — which can require rebuilding from scratch within months — the imperative is to build on open, modular standards such as Apache Iceberg and the Open Semantic Interchange, Pierson added.

“All of these things, being standards-based, are going help you decouple and modularize your approach so that when you do inevitably need to make a change, you can do that safely and swap it out with the next best thing,” he said. “That makes your solution 10 times better.”

Stay tuned for the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of Qlik Connect.

(* Disclosure: TheCUBE is a paid media partner for Qlik Connect. Neither Qlik, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE