AI Is Changing What It Means to Be an Expert

The first generation of AI adoption asked a simple question: What work can we automate?

The next question is harder: What happens to the people who were experts at doing it before?

Expertise has traditionally been built through practice. Engineers write and review code. Analysts work through problems. Marketers study audiences and campaigns. Architects design systems. Leaders accumulate judgment by making decisions and seeing what happens next. If AI increasingly performs those activities, where does that leave the role of the expert?

Members of the Senior Executive AI Think Tank—a curated group of leaders specializing in machine learning, generative AI and enterprise AI applications—are thinking about that transition from inside the organizations and disciplines being reshaped by AI. Together, they raise an important question that reaches into the design of the organization itself: How do you build a company where AI can do more of the work without losing the human expertise that tells the company what work is worth doing?

From Knowledge to Judgment

Pawan Anand, Portfolio Head for Communications, Media and Entertainment at LTM, sees expertise splitting into two categories.

“Codified knowledge will become cheaper as models absorb it, while tacit expertise, causal understanding and judgment under ambiguity will become more valuable because they determine where the model should be trusted, challenged or redirected,” he says.

This distinction matters because knowing information is increasingly different from knowing when that information applies.

“Experts will increasingly design decision systems, encode standards and teach the organization through feedback loops rather than personally execute every task,” Anand says.

For executives, that means expertise should be positioned where errors are costly and learning compounds. Instead of measuring senior specialists by how much work they personally complete, leaders can ask where their judgment has the greatest leverage.

“AI can produce the work, but an experienced person still needs to know what good looks like, what questions to ask and when an output simply doesn’t make sense.”

– Brock Murray, Co-Founder at seoplus+

SHARE IT

Keep Experts Close to the Output

Brock Murray, Co-Founder at seoplus+, believes expertise becomes more valuable when AI handles execution—but only if organizations redefine what they expect from experts.

“AI can produce the work, but an experienced person still needs to know what good looks like, what questions to ask and when an output simply doesn’t make sense,” he says.

That makes evaluation itself a core skill.

“Deep technical knowledge provides the context needed to use AI effectively,” Murray says. “AI-native companies will likely need fewer people focused purely on execution and more people who can oversee systems, connect different areas of the business and make informed decisions.”

For leaders, he recommends building cultures where experts are not simply asked to approve AI outputs. They should actively experiment with AI, test its limits and develop the curiosity required to adapt as tools change.

Calibrate the System, Not Just the Task

Maitrik Patel, Sr. Engineering Manager at Apple, argues that expertise becomes harder to replace precisely because execution is becoming easier to automate.

“When AI handles execution, the expert’s job shifts from producing work to calibrating the system that produces it,” he says.

Calibration requires a different form of technical depth. An expert must understand not only how to perform a task, but how to recognize subtle failures.

“Knowing how to do something well is not the same as knowing how to define what good looks like at scale, catch the failure mode the model has not seen yet, or recognize when a technically correct output is wrong for the context,” Patel says.

This is key when AI produces outputs that are technically plausible but wrong for the surrounding context.

“AI-native companies that understand this will organize experts around system calibration, not task queues,” he says. “The ones that do not will discover, too late, that they automated execution while quietly losing the judgment that made the execution worth trusting.”

Make Expertise More Leveraged

Pradeep Kumar Muthukamatchi, Principal Cloud Architect at Microsoft, sees AI expanding the reach of expertise rather than reducing its importance.

“As AI takes over specialized execution, experts shift from doing the work to defining goals, setting constraints, validating outcomes and making high-stakes decisions,” he says.

That shift can also change organizational design.

“In AI-native companies, organizations become flatter and more scalable,” Muthukamatchi says. “Small teams of experts, amplified by AI agents, can deliver the output of much larger organizations.”

From a leadership perspective, managers may spend less time allocating tasks and more time orchestrating human-AI systems, governance and decision quality.

“Those who win will be those who combine deep domain expertise with AI fluency to drive better outcomes at scale,” he says.

“Experts are valuable because they can set the standard, recognize when the system is outside it and decide when an exception changes the rule.”

– Rishi Katdare, Senior Technology Executive at Amazon Web Services (AWS)

SHARE IT

Organize Around Judgment Ownership

Rishi Katdare, Senior Technology Executive at Amazon Web Services (AWS), describes the transition as expertise moving “upstream.”

“The expert is no longer valuable because they can perform every task faster than anyone else,” he says. “They are valuable because they can set the standard, recognize when the system is outside it and decide when an exception changes the rule.”

That suggests a different way to structure AI-native organizations.

“AI-native companies should organize experts around judgment ownership rather than task queues,” Katdare says.

In practice, that means giving experts responsibility for standards, exceptions, workflow improvement and outcomes, not simply assigning them the hardest remaining tasks.

“The scalable asset becomes expert judgment embedded into how the company operates,” he says.

Shift From Functions to Outcomes

Venkata Kondepati, Manager of Data Architecture and Engineering at Ascentt, sees the expert role moving from production to definition.

“The expert’s role shifts from doing the work to defining what good work looks like,” he says.

That distinction also changes team structure.

“AI-native companies should, therefore, organize around outcomes, not functions,” Kondepati says. “Small teams can leverage AI to execute across traditionally separate specialties, while senior experts become architects, reviewers and decision-makers.”

For executives, the practical question is whether organizational charts still reflect how value is actually produced. If an AI-enabled team can execute across disciplines, rigid functional boundaries may become less useful than clear ownership of outcomes, quality standards and decisions.

“Leadership must focus less on managing production capacity and more on setting direction, designing systems, maintaining quality and knowing when AI should—and should not—be trusted,” he says.

Create Room for Emergence

Gabriella Goddard, CEO and Director of AI Catalyst Leader™ at Brainsparker Ltd, sees AI freeing experts to work on questions that extend beyond execution.

“As AI takes on more analysis, synthesis and execution, experts are freed to focus on judgment, discernment and developing those capabilities in others,” she says.

But she sees another possibility: AI can expand what experts are capable of imagining.

“AI gives experts richer raw material to think with: more perspectives, patterns and possibilities than they’ve ever had access to before,” Goddard says. “Combined with deep human expertise, this expands their understanding of the problem and can reveal opportunities they may never have considered before.”

She calls the resulting opportunity the “Emergence Era”—a space where human expertise and machine-generated possibilities combine to produce ideas neither would have found independently.

“I believe this will accelerate innovation dramatically,” she says. “The AI-native companies that lead won’t simply automate expertise, they’ll create the conditions where experts can make emergence part of everyday work.”

“With the rise of the internet, and now especially with the rise of AI, knowledge is very accessible to almost anyone.”

– Daria Rudnik, Team Architect and Executive Leadership Coach at Daria Rudnik Coaching & Consulting

SHARE IT

Broaden the Definition of Expertise

Daria Rudnik, Team Architect and Executive Leadership Coach at Daria Rudnik Coaching & Consulting, argues that expertise has historically been valuable partly because knowledge was difficult to access.

“With the rise of the internet, and now especially with the rise of AI, knowledge is very accessible to almost anyone,” she says.

The expert therefore becomes someone who combines multiple forms of understanding.

“The expert is no longer the person who knows the most,” Rudnik says. “It is the person who can combine domain knowledge, AI capabilities and business context to produce better outcomes.”

That has implications for promotion and development. Organizations should increasingly reward people who can translate specialized knowledge into business decisions and orchestrate AI effectively—not only those with the deepest technical specialization.

“For AI-native companies, that would mean selecting, training and promoting experts who can create this kind of impact rather than simply having deep technical knowledge,” she says.

Preserve the Interpreter

Fabio Danze Montini, Investor and Owner of FDM Industrial Sales & Marketing SL, argues that technical depth remains important—but its purpose changes.

“AI does not eliminate expertise; it changes where expertise creates value,” he says.

As execution becomes increasingly automated, interpretation becomes more consequential.

“The expert of the future is not the fastest operator, but the best interpreter and decision-maker,” Montini says.

For industrial businesses in particular, that means leaders should resist measuring expertise by throughput alone. The ability to understand customers, interrogate assumptions and connect technical outputs to commercial reality may become more important than the ability to produce another technical deliverable.

Protect the Apprenticeship

Mani Padisetti, Founder of Almost Magic Tech Lab, identifies a risk that can be easy to miss when companies focus on AI productivity: Expertise has to come from somewhere.

“Much of that judgment develops through doing ordinary work, making mistakes and seeing the consequences,” he says.

If companies automate too much of the entry-level work, they may inadvertently remove the experiences that create tomorrow’s experts.

“AI-native companies should give experienced specialists responsibility for developing successors,” Padisetti says, “with protected time for juniors to solve selected cases before seeing the AI’s answer, then follow what happened in practice. Otherwise, a lean company may be relying on expertise it has stopped replenishing.”

What Leaders Should Do Now

• Put experts where judgment has the greatest leverage. Move experts from routine execution toward setting standards, designing decision systems and guiding organizational learning.
• Make knowing what “good” looks like an explicit skill. Experts remain essential for evaluating AI outputs, asking the right questions and recognizing when an answer does not make sense.
• Use experts to calibrate AI systems, not just review their work. Give experts responsibility for defining quality, identifying unseen failure modes and determining when technically correct outputs are wrong for the context.
• Design organizations to amplify expertise. Small teams of experts can accomplish more when AI agents handle specialized execution, allowing leaders to focus on orchestration, governance and decision quality.
• Give experts ownership of exceptions and outcomes. Replace task queues with judgment ownership, with experts responsible for setting standards, reviewing exceptions and improving the systems they oversee.
• Organize teams around outcomes rather than traditional functions. Use AI to bridge specialties while positioning senior experts as architects, reviewers and decision-makers accountable for results.
• Create space for experts to use AI for discovery, not just productivity. Give experts room to combine AI-generated perspectives and possibilities with human creativity to uncover opportunities neither might find alone.
• Broaden the definition of expertise when hiring and promoting. Organizations should value people who can combine domain knowledge, AI capabilities and business context rather than relying solely on deep technical specialization.
• Value interpretation as much as execution. Measure experts by their ability to question assumptions, connect disciplines and make sound decisions under uncertainty—not simply by how quickly they produce work.
• Protect the apprenticeship that creates future experts. Actively develop successors and give junior employees opportunities to reason through selected cases before seeing AI’s answer.

The New Value of Expertise

AI is making specialized execution increasingly abundant, but it may be a more complicated transformation than simple substitution. Expertise is moving upstream—from producing answers to defining standards, interpreting outputs, managing exceptions, exercising judgment and developing the next generation of experts.

For AI-native companies, that means organizational design must evolve alongside technology. The focus should no longer be on who has the deepest technical knowledge, but on who can turn knowledge into reliable decisions, teach an AI-enabled organization what good looks like and remain accountable when the system encounters something it has never seen before. As AI takes on more of the work, those capabilities may become the foundation on which scalable expertise is built.