LinkedIn·Tuesday, 11 August 2026·16d ago
Most companies have their AI budget backwards. They picture the spend as models, infrastructure, and talent. Data and governance become the…
Felix Van De Maele
Entrepreneur, Founder, CEO of Collibra
Most companies have their AI budget backwards.
They picture the spend as models, infrastructure, and talent. Data and governance become the overhead you trim to free up cash for the exciting part.
It is exactly the reverse.
AI depends on the foundation underneath it: the data it retrieves, the governance that controls what it can access, and the systems of record it relies on. Yet those are often the same investments teams are tempted to raid to fund AI.
That is why I push back when the conversation becomes AI versus everything else. Cutting the foundation to fund the AI sitting on top of it doesn't get you very far.
If your AI is hallucinating, stuck in pilot, or surfacing data it should never see, spending more on the model won't fix the underlying problem.
Good discussion with Madeleine Streets at InformationWeek on how AI is changing the IT portfolio. Thanks for including my perspective.
https://lnkd.in/ekvmdwit
IT Strategy recent news | InformationWeek
https://www.informationweek.com/it-strategy/rethinking-the-it-portfolio-and-budget-in-the-ai-era
♥ 40💬 9↻ 1
View on LinkedIn Cross-referenced
Related on the wire
Agents are moving faster than the data and context they depend on, a problem I discussed with Informa TechTarget's Scott Thompson. Without…
More context isn't the win. Governed context is. There's a gap most AI strategies never name: the distance between the data a model can see…
I spend a lot of time talking to CIOs and CDOs. Most have an AI policy. Far fewer can tell me which production models are running without a…
If a company like Anthropic only achieved a 21% accuracy rate on its own data without the right business context, what do you think happens…
The world changed, and our category had to follow. For years, enterprise AI was defined by model capability. Today, the models are no…
The failure point in enterprise AI is rarely the model. It is the context around it. Ask two AI applications the same question about…