Every Function Has an AI Strategy. Who Owns the Work?

This week, I sat with a group of learning and transformation leaders to discuss how AI is changing capability development. One comment stayed with me.

The conversation is no longer: what skills do people need to learn?

Increasingly, it sounds more like: how do people use AI alongside the skills they already possess?

A project manager is no longer simply learning project management. They’re learning how to manage projects with AI. An analyst isn’t just learning analysis. They’re learning how to analyze with AI. A marketer isn’t simply learning marketing. They’re learning how to create, test, and optimize with AI.

That shift matters. Because it reveals a much larger transition underway within organizations.

The conversation is moving from “what work do humans perform?” to “what work belongs to humans, what belongs to AI, and what happens in between?”

That is not a learning question. It is a work design question.

Every function has an AI strategy

Across the enterprise, teams are launching initiatives to understand how AI may reshape their function. Marketing has initiatives. Finance has initiatives. HR has initiatives. Technology has initiatives. Operations has initiatives.

This is not a bad thing. Organizations should be experimenting, learning, and moving. But there is a question I rarely hear asked:

Who owns the work that sits between the functions?

Because work rarely happens inside a single function. Work moves through functions. Through decisions. Through handoffs. Through dependencies. Through informal expertise. Through judgment calls that never appear in a process map.

AI strategies I see often aim to optimize functions. Organizations succeed or fail at the systems level.

The foundational work is being delayed

At the same time, many organizations are quietly delaying investments in the very capabilities that make AI transformation possible:

• Talent intelligence
• Work architecture
• Process visibility
• Capability mapping
• Skills intelligence
• Expertise discovery

The logic is understandable. Business pressure is immediate. Optimization mandates are real. Leaders want movement. The assumption becomes: we’ll build the foundations later.

But later is usually when organizations discover they needed them in the first place.

Recently I sat in a conversation about workforce intelligence infrastructure and capability data. Everyone in the room agreed that understanding skills, experience, expertise, and capability would become increasingly important in an AI-enabled organization.

The challenge wasn’t belief. The challenge was prioritization. AI programs had sponsors. Capability infrastructure did not.

The organization wasn’t deciding whether visibility mattered. It was deciding whether visibility could wait. The answer was: not now.

The decision was rational. It may even have been correct given the constraints. But it revealed something important:

Organizations often invest in changing work before they invest in understanding it.

AI accelerates what already exists

AI will arrive. The question is whether organizations understand their work well enough to benefit from it. Because AI deployed into poorly understood workflows does not create transformation. It accelerates existing conditions.

Good work improves faster. Broken work breaks faster. Clear processes become more efficient. Invisible processes become more fragile. Local optimization becomes enterprise friction.

The spend is rarely the problem. The unmapped work usually is.

The capability question

There is another risk hiding underneath the technology conversation. Some of the work organizations are most eager to automate may also be the work that creates expertise.

The first draft analysis. The coordination work. The research. The routine decisions. The edge cases. The small problems that force people to think.

These tasks do not only produce output. They often produce judgment.

A senior engineer is usually a junior engineer who solved hundreds of small problems. A trusted operator is often someone who made mistakes, received feedback, and built pattern recognition through repetition.

Expertise rarely appears all at once. It accumulates through work.

The organizations that win the AI era will not simply ask: can AI do this work? They will also ask: what capabilities are formed by doing this work?

Because the question isn’t whether AI can perform the task. The question is: where will tomorrow’s experts come from once the work that created them disappears?

The six Execution Visibility questions

Before deploying AI at scale, leaders should understand:

• Where is value actually created?
• Which work creates expertise?
• Which work develops judgment?
• Which handoffs create friction?
• Which work should AI perform?
• Which work should remain human?

Those are not technology questions. They are work architecture questions.

Closing

The organizations that win the AI era may not be the organizations that deploy the most AI. They may be the organizations that understand their work better than their competitors.

Because AI changes tasks. But organizations still compete through judgment, expertise, and execution. And those things are still built through work.

See the real work. Then redesign it.

Originally posted on LinkedIn.