The Data Foundation Problem: Why “Bring AI to Your Data” Beats the AI Data Lake

Episode: Utilizing AI Episode 36: “Building a Trusted Data Foundation for Scalable Enterprise AI, Presented by OpenText”
Guests: Waqas Ahmed, VP of AI Development, OpenText (episode sponsor) · Keith Kirkpatrick, VP & Research Director, The Futurum Group
Host: Stephen Foskett, President, Tech Field Day, The Futurum Group
Episode Published: July 22, 2026

Listen: YouTube | Techstrong | Podbean

The Take

Most enterprise AI pilots don’t stall on the model. They stall on a data-architecture instinct that feels responsible but isn’t: consolidate everything into one data lake, then point the AI at it. Ahmed’s counter-argument, built from OpenText’s content- and service-management deployments, is that consolidation is exactly what breaks an AI system; it strips out the security, governance, and business-process context that made the data trustworthy in the first place. The fix isn’t more data in one place; it’s AI that can selectively reach the right data, wherever it already lives, without breaking the rules that governed it before AI showed up.

What You’ll Hear

Why the “just export it to a data lake” request that clients ask for first is the request that OpenText will push back on.
The one architectural principle — bring AI to your data, not data to your AI — that reframes every governance conversation in the episode.
A loan-processing story where the bottleneck was never the data itself, but who was allowed to see it and when.
The vulnerability-remediation numbers that show what “trust at scale” looks like once it leaves the demo.
Why Ahmed insists AI elevates jobs rather than eliminates them, and what that has to do with software developers today.

The Insights

Bring AI to Your Data, Not Data to Your AI

Ahmed’s framing is a direct rebuttal to the default enterprise instinct of centralizing data before applying AI to it. Aggregating everything into one store creates a context-management problem, a security problem, and an answer-reliability problem simultaneously, because the metadata, access rules, and business-process meaning that lived inside the original system of record don’t travel with the export. The implication: teams that treat data consolidation as an AI prerequisite are solving the wrong problem before they’ve started.

“You need to bring AI to your data and not data to your AI.” — Wakas Ahmed, VP of AI Development, OpenText

Information Access Is Not Consolidation — It’s Selection

Enterprises should separate “having data” from “having the right data at the right moment.” Ahmed argues that real trust comes from intelligent, dynamic selection — drawing relationships among clients, users, and documents on demand — rather than from a static lookup against a single aggregated table. Kirkpatrick’s insight from Futurum’s own research backs this: a fact can be accurate yet still irrelevant to the scenario, which he flags as a leading cause of stalled AI pilots.

“You cannot drive solid business outcomes unless you have all of those elements.” — Keith Kirkpatrick, VP & Research Director, The Futurum Group

Trust Only Counts Once It Survives Production

Demos succeed because someone assembled the perfect conditions for one run. The panel’s bar is higher: outcomes that hold up at scale, tracked over time. Ahmed points to OpenText’s Fortify security-remediation tooling, which cuts vulnerability triage from days to about an hour per issue, and a manufacturing client’s AI-assisted ticket routing running at 97% first-time accuracy globally — evidence, in his framing, that governed data access produces repeatable results rather than one-off wins.

“AI does not eliminate jobs — it elevates your job.” — Waqas Ahmed, VP of AI Development, OpenText

The Big Picture

The organizations most exposed here aren’t the ones lacking AI ambition — they’re the ones already mid-migration into a single data lake, treating consolidation as the AI-readiness milestone. Ahmed’s examples cut the other way: the financial-services client that scaled loan processing without ever centralizing its documents, the manufacturer running ticket routing at 97% first-time accuracy without collapsing its systems of record, the security team fixing issues in an hour instead of days because the context traveled with the data instead of being flattened out of it. Enterprises that export everything into one context-free store are building the exact fragility they’re trying to engineer away.

Listen & Resources

Listen to the full conversation: YouTube | Techstrong | Podbean

Mentioned in this Episode

OpenText solutions: Content management, Service management, Fortify Security Remediation Aviator
OpenText legacy asset base referenced: HP, Documentum (DELMC)
OpenText client insights (unnamed): Large financial-services firm (loan/document processing); large industrial manufacturing firm (AI-assisted ticket routing)
Upcoming OpenText World 2026 Event (October)

Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.

Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.

Other Insights From Futurum:

Can DataRobot’s Unified AI Governance Break the Silo Trap for Enterprise AI?

Going Beyond the Data Graveyard With Google’s Agentic Data Cloud as the New Semantic Core for Agentic AI

CIOs Consolidate Platform Spending as AI Moves From Pilot to Revenue Engine