Your "AI Strategy" Is Probably a PowerPoint Slide. That's the Problem.

Your "AI Strategy" Is Probably a PowerPoint Slide. That's the Problem.

Ask most CEOs who owns AI at their company, and you'll get a confident answer.

Ask what that person has actually shipped, governed, or shut down in the last 90 days, and the confidence usually disappears.

That gap is not an accident. It is the default outcome of treating AI as a topic to discuss in board meetings rather than a system to be led. And it is now the single most expensive blind spot in mid-market and private equity-backed companies, across healthcare, industrial operations, and logistics.

The numbers are not subtle. Companies have poured $30 to $40 billion into enterprise generative AI, and MIT's research found that 95% of organizations are getting zero measurable return. Separately, S&P Global Market Intelligence found that the share of companies abandoning most of their AI initiatives before production jumped from 17% to 42% in a single year, with the average organization scrapping 46% of its AI proof-of-concepts before they ever ship. Gartner expects over 40% of agentic AI projects to be canceled by the end of 2027 — not because the models failed, but because of "escalating costs, unclear business value or inadequate risk controls."

Read that last phrase again. That is not a technology problem. That is a leadership problem wearing a technology costume.

The Old Fractional CTO Job Is Already Gone

For years, a fractional CTO was hired to be a smart voice in the room: review the architecture, sanity-check the vendor, sit in on the board deck, offer a point of view when asked. High-level advising. Useful, but optional in most weeks.

That version of the role does not survive contact with AI.

The fractional executive market itself has said this out loud. As one 2026 analysis of the fractional CTO role put it: "The most significant shift is scope. A fractional CTO engaged in 2026 is frequently expected to own AI strategy, not just comment on it." And on what actually separates a credible operator from a talking head: "There is a meaningful difference between a leader who can discuss large language models in general terms and one who has built a procurement process around them, managed a failed implementation, or integrated an AI layer into production." (Fractionus)

That is exactly the line JHG operates on. Consultants advise. JHG leads. That means joining the executive team to deliver decisions and results, not a slide deck of recommendations someone else has to execute.

If your fractional CTO's AI contribution is a slide with a maturity curve on it, you have hired a narrator. You needed an operator.

Why This Keeps Happening: Nobody Owns the Outcome

The research is remarkably consistent on the root cause, and it is not talent, budget, or model quality.

KPMG's Global AI Pulse found that 75% of CEOs say they actively own AI as a strategic priority — but only 24% of organizations actually name the CEO or executive committee as accountable for AI-driven decisions. Where accountability is clearly defined, established ROI runs at 14% versus 4%. That is not a marginal improvement. That is more than three times the return, and the only variable that changed was ownership.

EY's Technology Pulse Poll found that 52% of department-level AI initiatives are running with no formal approval or oversight at all, and 78% of leaders admit AI adoption is outpacing their organization's ability to manage the resulting business risk. EY's own framing is worth repeating to any leadership team that thinks speed is the goal: "That's the velocity paradox leaders are navigating today, balancing urgency with accountability."

Meanwhile, the vacuum does not stay empty. A CIO/BlackFog survey of 2,000 workers found that 49% have adopted AI tools without employer approval, and 51% connected AI tools to work systems without IT's knowledge. Nearly a quarter are feeding company financials into unapproved tools. And two-thirds of the presidents and C-suite members surveyed appear to be fine with it. Leadership is not just failing to govern AI — in many companies, leadership is the source of the ungoverned behavior.

This is the pattern JHG sees inside client after client: plenty of AI activity, almost no accountable AI leadership. Pilots. Chatbot experiments. A department that quietly signed up for a tool. None of it connected to a person who can be asked, directly, "what did this produce, and what did it risk?"

Governance Is No Longer Optional, and It Is No Longer Abstract

If "AI governance" still sounds like a compliance checkbox to your leadership team, the market has already moved past that assumption, and it is moving fast enough to create real exposure.

Insurers are pulling out. Verisk's ISO has filed three new exclusion endorsements for generative AI across commercial general liability lines, and Berkley has already introduced an absolute AI exclusion across D&O, E&O, and fiduciary liability products — meaning a company using AI without a defensible governance process may find it has no coverage if something goes wrong. AI-related lawsuits grew 978% from 2021 to 2025, and 137% in the last year alone.

Boards do not know what they don't know. Deloitte's global board survey found that 66% of boards still describe themselves as having "limited to no knowledge or experience" with AI, while NACD data shows AI jumped from 27% to 65% of public-company boards' top agenda issues in five quarters. Boards are anxious about AI and unequipped to oversee it, at the same time.

Private equity has already started pricing this in. Grant Thornton found that only 9% of PE firms are confident they could pass an independent AI governance audit within 90 days — and that "almost every deal now involves product and tech diligence." Their term for companies that claim AI maturity without substance: vaporware. Accordion's research found that 75% of portfolio companies arrive at close either unprepared or only partially ready for AI deployment, and fewer than one in ten PE firms has a fully operational AI Center of Excellence. Sellers who cannot substantiate their AI claims face, in Grant Thornton's words, "valuation risk, slower deals, and potential changes to deal structure and timing." Buyers are not paying for AI potential. They are paying for AI proof.

If your company is founder-led, PE-backed, or preparing for a transaction, "we're exploring AI" is no longer a neutral answer. It is a liability disclosure.

Healthcare and Industrial Operations Don't Get a Grace Period

For companies operating in healthcare or physical operations — the sectors where JHG spends the most time — the governance gap is not theoretical. It is operational and legal, right now.

In healthcare, courts are already making clear that AI does not dilute accountability. As one legal expert put it plainly: "Saying 'the AI got it wrong' generally isn't a defense." (Medscape) HIPAA violations tied to improperly governed AI tools carry penalties up to $1.9 million per violation category per year, and public-facing AI chatbots are not HIPAA-compliant by default. Yet Black Book Research's survey of 182 hospital leaders found that only 22% are highly confident they could produce a complete AI audit trail within 30 days if a regulator or payer asked for one — and a third cite unclear ownership among IT, quality, and compliance as the reason governance keeps stalling.

In industrial and operational technology environments, the stakes are even more direct. Federal guidance from NSA, CISA, and international partners states outright that large language models "should almost certainly not be used to make safety decisions" autonomously, and warns that model drift and data poisoning are acute risks when AI touches OT systems. (NSA, December 2025) In logistics, the framing from one supply chain analytics firm captures the difference well: "In most industries, a struggling or failed AI pilot means wasted budget and bruised credibility. In supply chains, the cost can be more immediate." (project44)

Advisory-only technology leadership was never built for this. Nobody manages HIPAA exposure or an OT safety envelope from a quarterly strategy memo.

The JHG Perspective

JHG does not treat AI as a slide in the strategy deck. It is treated as a governed, operated system with an accountable owner — because the data leaves no other credible option.

A fractional CTO engagement built for this moment includes:

• Owning the AI roadmap, not commenting on someone else's — defining the two or three use cases that actually carry business value and risk, and killing the rest before they consume budget and credibility.
• Building a practical governance framework, not a policy binder — the handful of decision rights, review gates, and audit trails that actually get used, sized to the business's real data and compliance maturity, not a Fortune 500 template.
• Closing the shadow AI gap — finding out what employees are already using without approval, and building the fastest safe path to bring it into the light instead of banning it into hiding.
• Making AI diligence-ready — documented ownership, auditable decisions, and a defensible answer when a buyer, lender, insurer, or regulator asks who is accountable for what the AI does.
• Operating inside regulatory and safety boundaries that do not move — HIPAA, OT safety principles, and emerging state AI liability laws are not suggestions, and a fractional CTO who cannot operationalize them is not actually leading AI, regardless of title.
• Giving the CEO and board a plain-language read on real exposure — closing the knowledge gap the data shows exists at nearly every level, from department heads to the boardroom.

That is a materially different job than "advise the board on AI trends twice a quarter." It is closer to running a second, smaller company inside the company — one whose only product is making AI decisions the business can defend.

Executive Questions Worth Asking This Week

Before your next AI update, board meeting, or PE check-in, ask your technology leadership these questions directly:

• If a regulator, insurer, lender, or acquirer asked us to produce our AI governance documentation in 30 days, could we?
• Who is personally accountable for every AI initiative currently running in this business — not which department, which person?
• How many AI tools are employees using right now that were never approved, and do we actually know the answer?
• Has our fractional or full-time technology leader shipped, governed, or shut down an AI initiative in the last quarter — or only discussed one?
• If our AI strategy were tested in due diligence tomorrow, would it hold up, or would it be labeled vaporware?

If more than one of those answers makes you uncomfortable, the issue is not your AI tools. It is that no one is actually leading them.

That is the job JHG was built to do.