The AI Question That Isn’t About AI

The AI conversation in most boardrooms has become familiar: where to deploy it, how to scale it, what to build first. Most of the airtime goes to the platforms, the use cases, the implementation paths. But across the leaders we work with, the AI work they are finding most useful is heading somewhere else entirely.

The most consequential AI conversations we are part of barely involve AI. They are about strategy questions the AI moment has surfaced and made impossible to defer: where the organization creates value, what a business or function is uniquely good at and what it is not, what is worth protecting in how the work gets done today, and what is worth letting go. Senior leaders have been carrying these questions for years without much pressure to answer them. AI has not created any of them, but it has made it a lot harder to keep them off the table.

Why AI surfaces what earlier technology never did

AI is different from past technology in what it touches. The systems that came before automated work that was already well-defined, the kind someone had already specified, structured, and written down. The work AI is now pressing on is different: the judgment work that has lived with senior people because it required context and discretion no one had ever reduced to a process. That work cannot be handed to a model until it is first articulated, and articulating it forces a harder question about what the business or function is actually for. Most organizations have deferred that question for a long time, and AI is putting it back on the table, whether they want it there or not.

Consider a Fortune 500 biopharma we have been working with on its portfolio planning function. The brief was never about where to deploy AI to speed up the existing work. The leader was already asking the harder question: as AI reshapes what insight generation and decision support look like, what does this function need to become, what value can only this function create, and what capabilities does it need to get there? These are questions about the business strategy itself, and AI is what finally put them on the table.

Where this is showing up

This pattern is showing up across industries. Take Moderna. Late in 2024, the biotech created a new combined role called chief people and digital technology officer, and promoted its HR head, Tracey Franklin, into it. The Wall Street Journal covered the move last May, framing it as another sign of AI changing how companies organize their workforce. Franklin told the Journal her work is to redesign the company role by role, asking which jobs still belong with people and which can be handled by AI and other technology. Moderna has now built more than 3,000 tailored GPTs through its OpenAI partnership, including one that Franklin said takes over the kind of work an entry-level HR analyst would normally do. As a result, the org chart keeps shifting: new positions are being added, others eliminated, others rebuilt. That is not an AI deployment exercise; it is a leader doing the strategy work on what the function is for in an AI-enabled environment.

Klarna is the cautionary version. In 2024, the fintech announced that its AI assistant was doing the work of 700 customer service agents, and the company was scaling back its outsourced support as a result. A year later, CEO Sebastian Siemiatkowski told Bloomberg that cost had been overweighted in the redesign and that quality had suffered. The company has since walked it back. It is bringing back human agents, and the function has been rebuilt around a split: AI for the routine inquiries, people for the situations that actually require judgment. The technology worked, but the strategy underneath it did not. Nobody had done the work of asking what the function was actually for.

Why this is good news for senior leaders

For senior leaders, this is good news, though it carries a cost for the ones who miss it. The leaders who treat AI as a deployment problem are paying for it: another year of pretending the strategy question can wait, even though that question is what actually determines whether any of the technology creates value. The leaders treating the moment for what it is are having the harder conversation now, and the qualities it rewards have not changed: a real understanding of how the work gets done, the judgment to see what creates value and what doesn't, and the willingness to challenge their own assumptions about where the organization is heading. None of that is being replaced by AI; if anything, it has all become more central.

We are not watching this from the outside. At a recent team retreat, we found ourselves in exactly the same conversation we keep seeing across the organizations we work with: what is the work that matters most for us to do now, what is worth protecting in how we deliver, what needs to change. The questions AI is forcing into the open inside large organizations are the same ones it is forcing on the people who advise them.

The bottom line 

The most useful AI work happening right now barely involves AI. It is the strategy conversation AI has dragged into the open: where the organization creates value, what its people are uniquely good at, and what is worth protecting as the work changes. The leaders getting the most from this moment are treating AI as a forcing function for the strategy work they had been deferring, getting clear on what their business or function is for before they decide what to automate. The technology question can wait. The harder one has already arrived.

Andrea Schnepf 

P.S.: Answering the strategy question is where this starts. Acting on it is harder, and that is the work we are taking up on August 19. Join us: Making the AI Vision Real.