What You’ll Find This Week

The word "innovation" is everywhere in corporate America. The definition keeps shifting.

In the early days of this newsletter, I wrote about how big companies drift from innovation toward efficiency. They find what works, they optimize it, and eventually stop asking whether they're building the right thing. It's one of the most consistent patterns I cover.

What I didn't fully anticipate was a newer version of that drift. Companies aren't just moving away from innovation. They're renaming it. Today, a new senior innovation role in tech or consulting is almost always an AI role. Organizations have decided to treat them as the same thing.

IBM surveyed 2,000 CEOs this month across 33 countries. 76% of their organizations now have a Chief AI Officer. A year ago, that number was 26%. The C-suite has never reorganized that fast around a single idea. That's a 50-point jump in a single year. It’s the fastest the C-suite has reorganized around a single idea in recent memory.

Almost nobody covering that stat is asking the obvious follow-up: what happened to the mandate that used to sit somewhere else?

Here’s what you’ll find:

Don’t Miss Our Latest Podcast

This Week’s Article

In 2025, 26% of major companies had a Chief AI Officer. Today, it's 76%. (via IBM) In a single year, the fastest-growing executive role in corporate America became the person in charge of AI innovation.

Nobody asked what happened to the person who used to own that mandate.

Search for "Director of Innovation" on LinkedIn and you'll find over 5,000 open roles. The title isn't dead — not by a long shot. But look inside the job descriptions and notice what the mandate actually is. In tech and consulting, "innovation" now means something specific: AI deployment. "Drive AI transformation." "Lead our AI innovation roadmap." "Own the enterprise AI strategy." The word stayed, but the scope changed dramatically.

Innovation, as a corporate function, has been quietly colonized.

The CAIO is the most visible symptom. The strategic mandate for AI now has its own C-suite owner, its own budget line, its own seat at the table.

The innovation function used to own that mandate.

The innovation title is still everywhere. But it's being flanked from above and redefined from within. New AI-titled strategic roles are capturing the executive attention and budget that used to flow to innovation leadership. Inside the surviving "innovation" titles, the job descriptions are changing to match.

This wouldn't matter much if AI deployment and innovation were the same mandate. But they're not.

A Director of Innovation at TCS earns between $199,500 and $260,000. The role is fully AI-focused. The word "innovation" in the title means AI. At Arizona State University, the Director of AI and Emerging Technology Innovation has AI explicitly in the title and the mandate. At Honeywell Aerospace, the Director of Innovation Lab and PMO is, in practice, an AI delivery manager: AI tool deployment, PMO governance, transformation programs. The innovation lab is an AI lab.

The pattern holds consistently across tech, consulting, and enterprise software: a new senior innovation title is almost always an AI title.

Now look at the sectors where the pattern breaks...

At the Marzetti Company, the VP of Culinary and Sensory Innovation works on food product development; AI is not explicitly part of the mandate. At Panda Restaurant Group, the Head of Operations Innovation covers store concept design and remodel processes; AI is not called out.

At BioNTech, the Director of External Innovations manages clinical collaborations, R&D pipeline deals, and external licensing. At Mayo Clinic, the Director of Solution Design and Innovation Engineering oversees AI and machine learning alongside microservices, cloud architecture, IoT, and edge computing. The mandate is shaped by what the work requires. In both cases, what the work requires is more than AI.

Same job title. Completely different scope. In tech and consulting, "innovation" has become a synonym for AI deployment. In food, pharma, and healthcare operations, innovation retains its full scope as a discipline in its own right.

Pharma's exception is structural. The underlying work is physical, regulated, and ultimately gated by a clinical trial no AI can shortcut.

The pharmaceutical industry has received over $8 billion in annual venture funding for AI drug discovery. There are 173 AI-discovered drug programs in clinical development worldwide. Not one has cleared FDA approval yet. That's not a failure so much as a result of a process where Phase III trials take years and fail roughly 90% of the time, regardless of what generated the initial candidate. The FDA's draft guidance on AI in drug development, issued in January 2025, requires a credibility assessment framework, lifecycle maintenance plans, and full transparency about model architectures. The regulatory clock does not compress because the discovery tool was better.

Major pharmaceutical companies now allocate 5 to 10 percent of R&D budgets to AI and digital initiatives. That reflects an accurate read of where AI fits in a pipeline that still requires chemistry, biology, clinical medicine, and regulatory navigation. The most advanced fully AI-designed drug in development, rentosertib, was identified by an AI system called PandaOmics for treating idiopathic pulmonary fibrosis and has shown clinical efficacy in Phase IIa trials. It is notable. It is not yet approved.

The labor market reflects the same pattern. Axial Search's analysis of 420 innovation job postings found that Life Sciences leads all sectors in innovation hiring at 18% of all postings, with Financial Services second at 17%. The sectors most committed to standalone innovation leadership are precisely the ones where innovation output cannot be reduced to AI deployment. Regulatory compliance and patient-centered methodologies require dedicated leadership that AI adoption cannot satisfy.

Indeed's GenAI Skill Transformation Index analyzed 2,900 common work skills across 53.5 million U.S. job postings. In software development, 81% of skills fall in the "hybrid transformation" zone: GenAI handles or co-handles the bulk of execution. In nursing and healthcare delivery, 68% of skills fall in "minimal transformation." The work requires physical presence, patient judgment, and human interaction that cannot be delegated.

When a company decides that AI deployment is the innovation mandate, everything outside that definition loses its budget line. When the innovation output is a drug that has to survive a clinical trial, the mandate cannot afford to collapse.

The AbbVie case makes the distinction precise. AbbVie runs multiple active innovation roles with no AI in the title. These sit alongside a Chief Business and Strategy Officer whose external innovation mandate has generated over $20 billion in deals since 2023. The Director of AI External Innovation and Partnerships is a separate, additive role, carved out specifically to scout AI companies that can advance drug discovery. AbbVie didn't rename innovation as AI. They added an AI layer to an innovation structure that was already there. The method serves the mission. In tech, the method has become the mission.

The CAIO problem is not the role. Companies need leaders who can manage AI transformation at scale. The problem is what happens when the CAIO absorbs the entire innovation mandate by default.

When a company creates a CAIO and folds "innovation" into that charter, the message is explicit: innovation = AI. The budget follows. The headcount follows. And the innovation work that doesn't produce AI deliverables gets cut, not because anyone decided it wasn't valuable, but because it doesn't fit the new mandate.

Business model reinvention, new market creation, cultural transformation, platform pivots: none of these carry an AI ROI by default. They're slow, difficult to measure, and expensive to staff. They're also the forms of innovation that are hardest to copy. AI makes it cheap to replicate a feature. It doesn't make it cheap to replicate a business model or a company culture.

You're not more innovative because you have a CAIO. You're more automated. Those are different things, and conflating them is expensive.

The companies and industries that haven't collapsed these categories share something: their innovation output can't be satisfied by deploying a better model. The FDA doesn't approve LLMs as drug candidates. A new restaurant format is not a generative AI output. The forcing function of a domain that requires genuine novelty keeps the innovation mandate honest.

Mastercard offers a useful counterexample from inside the tech economy. The company has a Chief AI Officer and a separate Chief Innovation Officer in Ken Moore. Moore's mandate covers new payments infrastructure, financial inclusion, cybersecurity, and emerging markets. The CAIO owns AI capability. Ken Moore owns building what doesn't exist yet. At Mastercard, those are not the same job.

For everyone else, the question is whether anyone in leadership is keeping that distinction alive, or whether "innovation" has become a synonym for "AI deployment" and nobody noticed.

The CAIO proliferation tells you that most of the work that used to justify a corporate innovation function has been absorbed, automated, or made cheaper by AI. Horizon scanning, competitive intelligence, rapid prototyping, trend synthesis: AI handles most of it now. The question it doesn't answer is whether that's all the innovation that matters.

The pharmaceutical industry, under regulatory pressure that AI cannot shortcut, has not been able to make that substitution. Their innovation mandate stays broad because it has to. They have $8 billion in AI investment, 173 drugs in clinical trials, and zero approvals from AI-discovered candidates so far. They know better than anyone that the method and the mission are not the same thing.

Every company has innovation work that AI can't do: the category that doesn't exist yet, the customer relationship that requires reinvention, the business model that requires changing how you make money. The CAIO is well-positioned to automate what you already do. The question is whether anyone was given the mandate to build what doesn't exist yet.

The innovation function didn't die. It got reassigned. Whether anyone noticed depends on whether you've asked what the CAIO doesn't own.

Share This