AI is no longer just the territory of engineers and data scientists. Increasingly, the most valuable use cases happen when business professionals – marketers, healthcare workers, financial analysts, and managers – use AI tools themselves.

That’s the central message of new research by Prasanna “Sonny” Tambe, professor at Wharton and Faculty Co-Director of Wharton Human-AI Research. His paper in Management Science, Reskilling the Workforce for AI: Domain Expertise and Algorithmic Literacy, which was made possible through funding from The Mack Institution for Innovation Management, shows that firms capture more value from AI when algorithmic expertise is distributed across domain experts rather than concentrated in IT departments.

“The future belongs to workers who are bilingual – fluent in their field and fluent in AI,” Tambe says.

Expertise + AI = Strong Results

Historically, technical knowledge in firms has been centralized in IT or specialized data teams. If a sales manager needed an algorithm to predict customer churn, or a doctor wanted to test a new diagnostic model, they had to request help from specialists. This setup created bottlenecks.

Tambe’s research argues that AI is different from earlier technologies because it is general-purpose and deeply intertwined with local, domain-specific knowledge. “AI tools often generate predictions or recommendations that require contextual judgment,” Tambe explains. “You can’t separate the algorithm from the field knowledge.”

That’s why organizations increasingly want professionals who can combine both. Think of a nurse who can use predictive analytics to anticipate patient needs, or a marketing professional who can deploy generative AI to personalize campaigns.

The Proof is in the Data

To test how workforce needs are changing, Tambe analyzed two large-scale datasets:

1. Job postings from Lightcast (2013–2016): capturing employer demand for skills.
2. Workforce profiles from Revelio Labs (2008–2021): showing how skills are distributed inside public firms.

These were linked to financial data from Compustat-Capital IQ, allowing Tambe to study not just hiring trends but also how the market values these workforce shifts.

The findings are striking:

• Algorithmic skills are spreading beyond IT. By 2016, only one-third of postings requiring algorithmic expertise were for IT roles. The rest were for business-facing jobs that combined technical literacy with domain knowledge.
• Domain experts with decision-making power are adopting these skills fastest. Finance managers, HR leaders, and healthcare practitioners increasingly list AI and data science abilities on their resumes.
• No-code tools accelerate this trend. Software that lets non-programmers use algorithms makes it easier for domain experts to adopt AI capabilities.
• Firms that distribute AI expertise create more value. Market data shows investors assign higher valuations to companies that combine AI investments with domain experts who can use them.

“AI is becoming like Excel,” Tambe notes. “It’s not just for specialists anymore, it’s a basic productivity tool across many jobs.”

No-Code? Better Adoption.

A major driver of this shift is the rise of no-code and low-code tools. These platforms, ranging from visualization dashboards to generative AI assistants, remove the need for specialized programming knowledge.

“When the cost of using these tools falls, adoption spreads quickly,” Tambe says. “Suddenly, people who never thought of themselves as technical are doing data analysis or prompting AI systems in their daily work.”

This democratization has profound organizational effects. It allows firms to push algorithmic decision-making closer to the front lines, empowering workers who understand customer needs, patient contexts, or financial markets to adapt AI tools in real time.

The Market Rewards Decentralization

One of the most novel parts of Tambe’s research is linking these workforce shifts to financial outcomes. By connecting skill adoption data to firm valuations, he shows that companies don’t just adopt AI, they build intangible assets when they train domain experts to use it.

The results are clear: firms that combine algorithmic investment with a decentralized workforce of algorithmically skilled domain experts see higher market valuations.

Investors appear to recognize that this mix creates durable organizational advantages. Competitors can buy the same AI software, but they can’t easily replicate a workforce where decision-makers are already trained to integrate algorithms into their judgment.

Managerial Implications: Build a Bilingual Workforce

For business leaders, the message is urgent. AI investments alone won’t drive value. Firms must simultaneously reskill their workforce to spread algorithmic literacy across roles.

Tambe outlines three key implications:

1. Train domain experts, not just data scientists. “You don’t want AI locked in a back office,” he says. “The real gains come when the people making decisions can work with these tools directly.”
2. Leverage no-code platforms to accelerate adoption. Managers should actively evaluate which tools make it easiest for employees to experiment with AI.
3. Redesign workflows and decision rights. When domain experts gain AI capabilities, they should also have authority to act on insights without waiting for technical approval.

This shift isn’t simple. Training takes investment, and organizations need new governance structures to balance accountability with empowerment. But the payoff, Tambe argues, is significant: “Firms that make these changes are positioning themselves to lead in the AI economy.”

Bottom line

Tambe’s research underscores a simple but powerful truth: AI’s value is unlocked when the people closest to business problems know how to use it.

Reskilling the workforce isn’t just an HR initiative, it’s a strategic imperative. Companies that equip their domain experts with algorithmic skills will not only make better decisions but also build organizational assets that competitors can’t easily replicate.

Or as Tambe puts it: “AI is a general-purpose technology. To use it effectively, you need a general-purpose workforce, people everywhere in the organization who can speak both the language of their domain and the language of algorithms.”