Last week, I joined a meeting between a partner and one of their enterprise clients. They were working on their AI and data strategy, focused around governance and breaking down silos to make data accessible across the organization.
I asked a question I’ve been hearing a lot lately: How much pressure are you getting from leadership and the board to deploy AI and prove ROI?
In many meetings, that question gets an exasperated sigh. But their answer surprised me.
“Not so much on AI,” they said. “The real pressure is showing ROI on our data strategy.”
The Disconnect Between Data and Outcomes
“Driving ROI from Data and AI” sounds great in an executive briefing. But as we dug in, it became clear what executive leadership actually wanted was much simpler: make this data produce tangible value.
That’s a disconnect. When we mapped out how that value was expected to materialize, there was no operational bridge. The data strategy was being treated like an isolated silo, as a standalone function expected to produce returns on its own balance sheet. Exactly how that was supposed to work was never defined.
This idea skips the most critical principle:
A data strategy doesn’t create value in isolation. It creates value by driving change in other parts of the organization.
Where ROI Actually Lives
We were discussing customer data, which is my wheelhouse, so the conversation naturally turned to where that data should actually move the needle:
Go-to-Market & Marketing: Which customer segments warrant investment, which campaigns drive return, and which channels convert.
Customer Support: Clarifying where self-service genuinely solves problems versus where to invest in people.
Resource Allocation: If you have an incremental dollar to deploy, does it belong in outbound GTM, frontline support capacity, or self-service infrastructure?
That last trade-off is ultimately an operating decision, not a data architecture decision. Yet that is precisely the type of cross-functional decision good data should inform, and it is the only place where the financial return is actually captured.
The Accountability Trap
To their credit, this leadership team wasn’t falling into the trap of spending aggressively on speculative AI just to show momentum, for instance the sort of silly “tokenmaxxing” we saw earlier this year that burned through too many budgets with little to show for it and gave “ROI on AI” a bad name. This team’s appetite for operational connection is there; it just lacked connective tissue.
Across the industry, organizations are hunting for ROI but haven’t internalized where those returns actually originate. They don’t appear on the data team’s scorecard. They show up in marketing, sales velocity, customer retention, supply chain resilience, manufacturing throughput, and working capital management.
If models and analytics don’t change how those line functions operate day to day, you aren’t running a data-driven enterprise, you’re just hosting an expensive data repository.
It sounds obvious, but the gap between acknowledging this and executing on it is where too much capital quietly evaporates.
Today’s data and AI leaders are trapped in an unfair dynamic: they’re held accountable for driving bottom-line value without the cross-functional mandate to direct the business units that actually generate those returns. That is an organizational structure built to disappoint everyone involved.
The answer isn’t another isolated data roadmap. It’s a business operating strategy that the data architecture is explicitly engineered to serve, run by business unit leaders who are on the hook to act on what the data reveals.
Your data strategy supports the business strategy. Not the other way around.
Where We Go From Here
I’ve had numerous conversations like this recently, and the pattern is unmistakable. There is a fundamental architecture gap between data investment and operational realization, and it’s an area we’re actively exploring through deeper research.
And of course there’s the huge overhang of skepticism and fear about AI, with the flames being fanned on a daily basis.
I choose to be an optimist; while there is risk, the upside is simply too high. We’ll soon be kicking off a new initiative focused on showcasing the operational value of AI and data. Our first deliverable launches in early November—more details coming shortly. Stay tuned.