CEOs: Are you setting your Head of AI up for success or failure?
Most CEOs genuinely want AI to succeed, but don't realize how much they're responsible for. Here's how CEOs and Heads of AI (CIOs, CAIOs) can collaborate to improve ROI.
“You need to understand AI strategy because these are leadership decisions that I can’t make for you.” - VP Enterprise AI
Earlier this week I had a chat with the Head of AI at a B2B company. She described a business where there’s no strategic clarity, little cross-functional collaboration, and AI keeps getting stuffed back into IT because that’s where tech belongs, right?
Leadership looks to her to deliver ROI from AI, but aren't creating the conditions that would make that success possible.
AI can’t be added to business as usual
Her company isn’t unique, and these issues help explain the lack of ROI from AI. Only 6% of companies have achieved over 5% improvement on EBIT from AI (McKinsey*), and just 20% of companies are capturing 74% of all AI-driven value. (PWC*).
There are plenty of reasons why companies aren't seeing results, but they all share a root cause. AI can’t be added to an outdated "business OS", ie. the underlying logic that informs how your business operates.
We inherited the old operating system from Ford and his famous assembly line over 100 years ago. We love putting every function in its own box, drawing neat hierarchies, and rewarding senior people with their own fiefdoms they can control.
AI blows all of this up.
AI is a horizontal transformation – it doesn’t sit in any single function, and touches all of them. It eliminates entire coordination layers that were built for humans and our slow, messy ways of getting things done. And it turbocharges whatever it’s added to, multiplying either value or complexity debt.
In a B2B software company’s Order-to-Cash process, for example, AI agents can now qualify prospects, draft contracts, onboard customers, generate invoices, and chase payments. When connected end-to-end, they wipe out the handoffs, status meetings, email chains, and escalations that patched the holes between sales, legal, finance, customer support, and collections. One high-value move decreases costs, increases velocity, and improves CX and loyalty.
But if you haven’t redesigned the entire process first, the small things that humans once caught now race through the entire chain at the speed of light. Fixing the consequences is more expensive than your original AI investment.
“70% of AI roadblocks are caused by people, organization and processes.”
- BCG, Build for the Future 2025
Redesign starts with the CEO, not AI or operations.
AI requires a fundamental rethink of the business. Yes, we’re talking workflows, which you might assume belong in the COOs or Head of AI’s remit. But workflow problems are symptoms of incentive systems, power structures, ownership models, and priorities -- all CEO responsibilities.
Transformation is impossible when the structure and incentives reward the old way of working. P&Ls are where power sits; if they're aligned to functions, then there’s no incentive for anyone to play ball. The Head of AI walks into that world with no P&L of their own and is somehow expected to deliver enterprise-wide outcomes.
Functional P&Ls also have a huge impact on how AI success is measured. Traditional finance teams, locked into their chart of accounts and functional cost centers, can easily report “we saved $X on headcount with the AI chatbot,” but miss the downstream impacts like rising customer churn or damaged loyalty that destroys more value than was saved.
There’s a reason why the highest-performing AI-native companies are organized around solutions or customer segments. Forrester and others have spent years documenting the financial benefits of organizing around customers rather than functions. AI simply raises the stakes and forces the issue.
For AI to pay off, the CEO needs to shape the conditions for success: which means aligning talent, capabilities, P&Ls, data, metrics – basically everything – around your priority outcomes.
“When functions evolve independently, performance improves incrementally. When they’re redesigned as an integrated system, performance improvement compounds…. CEOs who are actively redesigning how cross-functional teams work together are more than twice as likely to have delivered on their business objectives.”
- IBM, Rewiring the C-Cuite, 2026
The real division of labor
In the dance of AI transformation, the CEO leads and the head of AI follows and supports. When business and technology in lockstep, you can evolve the organization into a fast, efficient value creation machine.
Here’s how that breaks down:
When the left column is vague or missing, the right column becomes an impossible job no matter how capable the person in it is.
“Organizations with CEO‑led, enterprise‑wide alignment are 2.5x more likely to achieve significant value from Generative AI.”
– Accenture, Making Reinvention Real with Gen AI, 2025
What’s next for CEOs and Heads of AI
CEOs, it’s time for some honest self-reflection. Here are four questions you can ask yourself and others in your organization.
Clarity: Have you identified the 1 - 2 most important business outcomes AI should drive? Should revenue growth come from current or new customers? Should profitability from higher-value experiences, lower cost to serve, or reducing complexity? Are you driving innovation or protecting your core? Your strategy should make it easy to prioritize a small set of high-value AI initiatives.
Coherence: Are you focused on outputs or outcomes? Flip the focus from what everyone does to what outcomes we can create together — and how we measure that integrated success. Same thing at the initiative level: you’re leaving money on the table if capabilities, talent and data aren’t all creating a compound flywheel effect towards a clear outcome.
Complexity: Is it easy to get work done? Find a couple mid-level managers and ask them honestly: how many people, approvals, layers, and systems does it take to get something done? Which decisions are painful to make? This is the work to be redesigned before adding, or expecting ROI from, AI.
Are you doing your part? Ask your Head of AI directly what’s in their way. If the answer involves things outside their authority, that’s your list. It probably includes budget control, direct CEO reporting, and help with greasing the cross-functional skids, among other things.
“It’s not the technology that is the limiting factor. We are the limiting factor.”
– Jacek Olczak, Group CEO, Philip Morris International (PMI), Switzerland
What Heads of AI can still do
Even without perfect CEO air cover, Heads of AI can still make progress.
Force prioritization. If leadership wants to pursue twenty AI opportunities simultaneously, force a commitment to the most important business outcome — then use that as a prioritization criteria to remove misaligned pilots and programs.
Solve cross-functional complexity from the bottom up. Find the leaders who are stuck on the same cross-functional bottlenecks — like onboarding, quote to cash, etc. — and make AI the excuse to redesign the workflow together. Prioritize the journeys that cause both customer and employee/business pain, giving you twice the impact for the same work.
Build active, evidence-based leadership. Work with the finance team to create an integrated way of measuring the total financial impact of AI pilots and programs; then test different types (ie. point solutions versus cross-functional workflows) so that you have the evidence for additional funding and the deeper business overhauls that are needed.
Cultivate non-technical skills. Your role is more about influence, collaboration, and evidence-building, which requires deep fluency in the language of business and people. Lead with what the tech means for everyone, not the tech itself.
Bottom line:
AI is a horizontal business transformation, not another tech program. The CEOs who understand that distinction will set their AI heads up to succeed at delivering outsized value.
Take the free assessment
Find out exactly where your gaps are with the Possibility Playbook Foundations assessment. It’s a lightweight, free version of our more robust maturity model. Unlike other assessments, this is intended for CEOs who know they need to redesign the business but are unsure what it looks like or where to start.
P.S. I help mid-market, complex businesses like B2B bridge the gap between business and AI to boost ROI. AI is my third horizontal transformation, after customer centricity and sustainability, and they all share the same shift to a new, more integrated OS.
If this article resonated, I'd love to hear where you're seeing friction between AI ambition and organizational reality. Feel free to comment below or reach out directly to learn more.
Sources:
McKinsey, The State of AI report, 2025
PWC, Decoding ROI from AI, April 2026
IBM, Solving the AI ROI Puzzle, July 2025
IBM, Rewiring the C-suite: The fast track to 2030, May 2026
BCG: Where is the Value in AI? Oct 2024
