Topline Enterprise AI Usage Data

The report offers several levels of executive summary. Here is what interested me in the report (not in order of priority):

• Nearly three-quarters of executives surveyed claim that they already see a positive ROI. That doesn’t necessarily contradict the MIT study that said 95% of pilots failed to scale and deliver an ROI.

• Chief AI Officer (CAIO) roles are present in 60% of enterprises (technically executives representing enterprises, but too tedious to write each time). Remind of Chief Digital Officers, Chief Mobile Officers, and even Chief Metaverse Officers - yeah. The authors’ take - accountability for AI is in the C-Suite. My POV - likely a temporary point person (or choke point) to coordinate activities, investment, and strategy around a truly transformative technology.

• Knowledge and familiarity with GenAI in marketing departments went down by six percentage points year over year. Consistent with this perception, use of GenAI applications in Marketing / Sales is almost the lowest among all functions.

• The percentage of professionals who claim deep understanding of GenAI or “very knowledgeable” is shockingly high - above 70-80% in most industries.

• “Daily” or “Fairly Often” usage of GenAI tools is above 80-90% in most industries. GenAI is starting to feel a bit like the Internet or Broadband … we need to move beyond such macro frequencies to truly understand usage, comfort, and reliance on the applications.

• Business tasks with the highest adoption are not customer-facing - more augmenting of humans. Not a surprise, but the depth of data by task here is helpful.

• Use of GenAI to create content made some of the largest jumps from 2024 to 2025.

• ChatGPT (67%) and MS Copilot (58%) are top-used tools by employees. This chart is on page 32 of the report. It also shows “used in the past and do not plan to use again.”

• Human abilities to use GenAI tools now seem to be the bottleneck in deployment. That said, the research didn’t offer up much about data quality or some of the other more traditional barriers.

Context: Wharton published its annual report (PPT) on business use of AI. It is a 90-page PPT with generous amounts of data - not inexpensive. The results primarily focus on an executive survey of small-, medium-, and enterprise-sized businesses in North America (and beyond). The PPT offers data cuts by role, size of business, industry, role seniority, and more. The report is very comprehensive. Source: Wharton Human-AI Research and GBK Collective. 

While the data is good, I don’t recommend reading through the 90-page PPT. It offers some analysis (e.g., what grew/what shrunk) and (for those who know me well) what I would call a “data dump.” My best guess is that the primary researchers and authors will use more sophisticated cuts of the data in their own upcoming research and reports. My only other criticism of an otherwise very extensive report would be the hypothetical questions.