How Business Really Adapts to AI: 3 Mindset Shifts Leaders Can’t Ignore

Even as more than three-quarters of small businesses say they’re using artificial intelligence (AI)¹, many of them are in early experimental phases or automating basic tasks such as writing, searching, summarizing, analyzing, coding, and producing content more quickly.

But this approach merely scratches the surface, says Paul R. Carlile, the Peter and Deborah Wexler Professor of Management and Information Systems at Boston University’s Questrom School of Business. Leaders who assume automation is the extent of AI’s usefulness will eventually fall behind.

The larger opportunity, Carlile says, is using AI to detect problems earlier, improve quality, strengthen judgment, accelerate learning, redesign workflows, and help businesses move more effectively between creating value and capturing value. Here, Carlile outlines three mindset shifts business leaders must consider if they hope to maximize the full capability AI has to offer.

1. Beginning with the business problem rather than the AI tool itself

New AI tools arrive with impressive capabilities, naturally persuading business leaders to ask where a tool might be deployed. “This can lead them to automate work that should be redesigned, improve a task that is not the true source of the problem, or deploy a system that produces output without creating a meaningful outcome,” Carlile says.

At Questrom, students are taught that leaders should begin by asking:

• What consequential problem are we trying to solve?
• What is happening now and how is the current system producing that outcome?
• Why does it matter?
• Who is affected, and where and when does the problem occur?

“Only after the problem has been framed and diagnosed should leaders ask what type of insight, prediction, recommendation, creation, or action is needed—and then determine which AI capability and tool are appropriate,” Carlile says. Doing so shifts the focus from task efficiency to business value.

2. Moving beyond what work AI can replace

Rather than focusing on which tasks or positions can be replaced by AI, Carlile argues that leaders should design ways humans and AI can make one another more capable. “Much of the current conversation treats the relationship between people and AI as a contest,” he says. “A task is classified as either human or machine work. That framing is too narrow.”

At Questrom, for instance, Carlile says the goal is not to train future business leaders to supervise AI superficially but to “use AI while also questioning its outputs, testing its assumptions, evaluating its consequences, and deciding where human judgment must remain central.”

3. Transitioning from isolated gains to an adaptive value cycle

Carlile says most organizations are structured to do one of two things well: optimize and scale what already exists or explore and create something new. “The people, processes, measures, and cultures required for those activities are different,” he explains. “Improvement rewards discipline, repeatability, and reliability. Innovation requires experimentation, ambiguity, and a willingness to challenge existing assumptions.”

He recommends that leaders apply AI in a cycle that creates and captures value. The organization cycle taught at Questrom looks like this:

Improve → Escalate → Innovate → Operationalize → Improve again

“AI can accelerate learning at every stage by analyzing operational data, revealing anomalies and weak signals, generating possible explanations, simulating alternatives, and helping codify new ways of working,” Carlile says. “But leaders must determine when the organization should remain within the existing system and when it should move beyond it.”

Business leaders can use AI to increase productivity, standardize work, and substitute technology for labor. Or, with a few shifts in mindset, they can use it to build a truly better business. “You can design better problem-solving systems, develop human and AI capability together, move more effectively between improvement and innovation, and create more adaptive growth strategies,” Carlile says. “Technology will not make that choice. Leaders will and they remain accountable for the consequences.”

^{1 Survey: Small Businesses Embrace AI — But Need Training and Support to Fully Harness It (Goldman Sachs, 2026)}

Click here to learn more about Boston University’s Questrom School of Business.