Wharton report says AI for business is now delivering results

Business leaders have had good reasons to be skeptical of AI at work.

Early vendors overpromised, most tools were confined to narrow functions, and many executives had little hands-on experience with the latest models. As a result, numerous internal AI initiatives never advanced beyond experimentation, as an MIT report noted.

But a new study from Wharton Business School reveals that enterprise AI has rapidly matured. When it comes to large and medium-sized businesses, "roughly three in four already see positive returns," writes the Wharton researchers.

They also found that "four out of five [businesses] see Gen AI investments paying off in about two to three years."

So, the technical obstacles that slowed earlier AI adoption have mostly faded. We've seen AI models take remarkable leaps in agency, reasoning, and contextual understanding. (See "AI for Startup Leaders" for a deeper analysis of this progression.)

What remains are human barriers. Especially when AI is introduced to teams where people still act as the glue (and the bottleneck) between siloed systems. This is why, even though the success rate of internal AI projects has grown, adoption remains uneven across teams within organizations.

Developers are leading the way

AI usage differs significantly by function within a company. Wharton's findings highlight a clear trend where IT teams are furthest ahead, both in confidence and in daily use of AI tools:

"Daily use is common, with IT and Purchasing/Procurement out front while Marketing/Sales and Operations trail."

IT departments are also where organizations see the clearest measurable gains, especially with code generation. That's because coding is structured, system-oriented, and verifiable, which is an ideal environment for AI success.

By contrast, Sales adoption is flat and Marketing adoption has dipped slightly. Wharton warns that without cross-functional coordination, AI progress will stall outside IT. Not because the tools are incapable, but because companies struggle to help teams adopt them successfully.

Aaron Levie, CEO of Box, commented on the Wharton study and captured this dynamic precisely:

"The amount of output you get from AI agents will be directly correlated to how much you change (or reset) your workflow. It's a continuum from those designing a process from scratch … to those not willing to change a thing."

Some companies have already embraced this shift. Instead of reactively greenlighting AI experiments from individual innovators, they are rolling out AI access broadly. From Wharton:

"70% of firms allow all employees usage access… More Tier 2 and Tier 3 enterprises now allow 'any' employee to use Gen AI."

One major change in the latest study compares to earlier surveys is executive suite participation. Wharton notes that, "Executive leadership in Gen AI adoption has surged (67%, +16pp YoY)."

Three years ago, AI was driven largely by individual contributors and line-of-business managers. Now the C-suite is directly engaged. As executives experience personal productivity gains, they are accelerating AI access across their organizations.

Why cross-functional AI is the next frontier

Function-specific AI tools, like "AI for sales," "AI for marketing," and "AI for engineering" have not lived up to their promises of company growth. These tools improve team-level productivity but also reinforce the same organizational silos that slow organization down.

For example, Marketing and Sales could successfully adopt a GTM AI tool but still become stymied because their workflows rely heavily on data owned by Product, Finance, and Support.

What's needed is a shared intelligence layer that bridges these systems, so that connected teams can reach escape velocity together. To achieve this acceleration, leadership needs to help the company become AI native. This includes:

• Granting employees and AI agents shared access to relevant business data
• Educating employees about AI's reasoning and agentic capabilities
• Making AI accessible in the spaces where work is discussed and executed, such as in Slack channels

When teams collaborate through a unified intelligence layer that understands your product analytics, financial reports, support interactions, and sales data holistically, they are enabled to make better decision, faster.

Becoming AI native

Wharton's report ends with a warning:

"Gen AI usage has become mainstream… The divide that remains is cultural. Open access, faster rollout, and clearer guardrails are what put leaders ahead of laggards."

In other words, the companies who will see the biggest gains from AI are redesigning how work flows across teams. And that's where Adapt comes in.

Adapt helps leaders move beyond isolated AI tools and build shared, cross-functional AI workflows that connect teams around the same intelligence layer. Instead of relying on a few skilled employees to stitch systems together or report on data, Adapt turns AI into a powerful capability for everyone.

As AI becomes mainstream, the real advantage belongs to companies that operationalize it across the business.