The board has stopped applauding AI deployments. Now it wants to see return on investment.
That’s the central message from Tope Iluyomade, executive advisor on AI to CIOs and former dual CIO/CISO, in a recent episode of the AI Impact podcast hosted by Larriding CEO Russ Fradin. With more 20 years in enterprise technology, spanning robotics, Wall Street critical infrastructure, and Fortune 100 executive roles, Tope brings a practitioner’s clarity to questions that most technology leaders are still fumbling through.
Tope begins by putting AI in a class by itself: “Many people think of AI as a tool, but AI doesn’t fit into a stack. I’ve never seen a tool that continuously improves itself.” For Tope, this distinction carries real strategic weight. A hammer doesn’t get better. AI does. That means it doesn’t belong in the tool stack; it belongs in the decision loop. And that placement changes everything about how CIOs should govern it, budget for it, and explain decisions related to AI to their boards.
Tope places AI in an actionable framework that he calls the two-layer portfolio model: Foundational layer. This layer is made up of a core set of AI tools, including OpenAI, Claude, and similar platforms, that everyone in the organization can access.
Specialist layer. For niche tools, Tope advises CIOs to avoid annual licenses. Instead, he recommends deploying specialized AI tools on a project basis and requiring proof of outcomes before renewal.
This approach resists the fragmentation of employee efforts across too many, often overlapping, tools that results from letting a thousand flowers bloom.
“Last year there was a lot more flexibility in going to a CFO and the board to experiment with AI,” Tope says. “This year you’re going to find it’s more of a P&L conversation. What does this do to our financial outcomes?” The implication for CIOs is clear: the vocabulary has to shift from deployment metrics to economic value.
Tope, like other guests in our AI Impact series, is emphatic that the partnership between the CIO and the Chief Human Resources Officer is no longer optional. Russ adds his opinion: the CIO has evolved from being seen as “almost an impediment; I’ve got to get through these guys” to a genuine strategic partner.
Tope argues this transformation accelerates further when you factor in AI agents as workforce participants. “CIOs today manage more than just human operators; they’re managing AI operators as well.”
This is where the CHRO becomes essential. Tope describes an “AI flight school” model, in which CIOs and HR co-design enablement curricula, rather than just licensing tools, deploying them, and hoping adoption follows.
The goal is building workforce competency systematically, not accidentally. And with employees simultaneously anxious about job displacement and eager to stay relevant, Tope argues that cultural enablement is the only honest answer: “It’s not about AI replacing you. But those that are not using AI are going to be replaced by those that are.”
On the security side, Tope rejects both extremes: using security concerns as a brake on AI uptake, and recklessly innovating without concern for unwanted impacts. “It’s not about uncontrolled speed versus controlled speed; it’s about engineered speed.” Security guardrails built into deployment pipelines don’t slow engineers down; they enable engineers to move with confidence.
Tope has a nuanced opinion as to the value of data: “Data is not necessarily the value; it’s the cognitive intelligence on what to do with that data that lives with the company.” He suggests introducing data masking and tokenization at the point of AI ingestion, invisible to pipeline participants and the end user.
Tope’s take on SaaS is nuanced: “SaaS is not dead; it’s going to transform.” But he does see a binary outcome forming. Ecosystem players with deep integrations will survive and adapt. Commoditized feature vendors will face mounting pressure as AI coding tools lower the barrier to building in-house alternatives.
What accelerates that pressure is MCP architecture. Tope describes it as quietly reshaping every build-versus-buy decision in enterprise software. New entrants building solutions on MCP from day one can offer lower complexity, faster integration, and cheaper services, without the technical debt that encumbers legacy platforms.
For enterprise CIOs, the message is strategic: know which of your SaaS vendors are ecosystem plays and which are just selling features. The latter are more replaceable than they were 18 months ago, and the gap is widening.
Across the full conversation, Tope’s through-line is measurement: You can’t manage what you don’t measure. The CIOs who will lead in 2026 and beyond are those who can translate AI investment into economic language, govern it culturally before it fails technically, and work with HR, CFOs, and operational practitioners to make accountability real.
You can view the full episode on YouTube or reach out to Larridin for more information.