The Wharton School and GBK Collective’s third annual AI Adoption Report, titled “Accountable Acceleration: Gen AI Fast-Tracks into the Enterprise,” reveals that 2025 marks a turning point for artificial intelligence in business, moving from experimentation to measurable performance and ROI.

From Exploration to Everyday Use

According to the report, 82% of enterprise leaders now use generative AI at least weekly (up 10 points year-over-year), and nearly half (46%) use it daily. Adoption has spread beyond early adopters in IT and procurement to functions like HR, finance, and legal.

Top daily tasks include data analysis (73%), document summarization (70%), and content creation (68%) — practical, repeatable workflows that now underpin productivity across the enterprise.

Large enterprises have closed much of the adoption gap with smaller, more agile firms, while lagging industries like retail and manufacturing still trail sectors such as Tech/Telecom, Banking, and Professional Services, which report 90%+ weekly usage.

“The story has shifted from curiosity to competence,” the authors write. “Gen AI is now part of daily work.”

Measuring ROI and Shifting Budgets

A defining theme of 2025 is accountability. Nearly three-quarters (72%) of companies now track structured ROI metrics, focusing on profitability, productivity, and throughput, and three out of four leaders already report positive returns from Gen AI investments.

Wharton’s data show that 88% of enterprises plan to increase AI budgets within the next 12 months, with 62% expecting double-digit growth. Roughly one-third of those budgets are allocated to internal R&D, signaling a move toward building custom, enterprise-specific AI capabilities rather than relying solely on external tools.

Among industries, Tech/Telecom leads with 88% reporting positive ROI, followed by Banking and Professional Services (≈83%). Retail lags at 54%, still working to translate automation into bottom-line results.

“Budget discipline and ROI rigor are becoming the new operating model,” the report notes.

People, Not Tools, Are the Bottleneck

While technology adoption is soaring, Wharton’s researchers emphasize that human capital is now the decisive lever in achieving scalable success.

• 89% of executives agree Gen AI enhances employee skills, yet 43% warn of skill atrophy if training doesn’t keep pace.

• Chief AI Officers (CAIOs) now exist in 60% of enterprises, signaling C-suite accountability.

• Training investment, however, is slipping (–8 percentage points year-over-year), and confidence in upskilling programs has declined by 14%.

The result is a growing divide between empowered, AI-fluent teams and those constrained by cultural or skill gaps. Organizations are experimenting with blended strategies, combining employee-led training, new hires, and hands-on pilot projects, but many still lack the leadership and change-management capacity to scale those efforts effectively.

AI Agents Begin to Take Hold

New in this year’s study is a snapshot of how enterprises are adopting AI agents, which are autonomous systems that handle workflow automation, analytics, or customer support tasks.

Over half (58%) of decision-makers say their companies are already using AI agents, mostly for process automation, analytics, or internal support.

As one tech executive quoted in the report described:

“AI agents now triage internal tickets, monitor DevOps, and flag finance anomalies before humans even notice. They’re still supervised, but they’re freeing up hours daily.”

An Inflection Point Ahead

Looking toward 2026, the researchers predict a new phase: performance at scale.

Enterprises will move from “accountable acceleration”, measuring and managing AI ROI, to rewiring entire workflows and deploying agentic systems across departments.

Four out of five leaders believe Gen AI investments will fully pay off within two to three years.

As Wharton’s co-director of Human-AI Research, Professor Stefano Puntoni, summarizes,

“This isn’t about whether AI will transform work.. The question now is how fast organizations can align people, process, and governance to capture the full return.”

Key Statistics at a Glance

Bottom Line

The 2025 Wharton–GBK report confirms what many in enterprise technology already sense: the generative AI boom is entering its next phase; one defined less by hype and more by data, governance, and results.

For most companies, that means moving from experimentation to operational integration, and from “AI pilots” to measurable performance at scale.