Want Quick AI Adoption? Give It Harder Tasks
AI is improving faster than most companies can put it to use. The gap between what the tools can do and the value a business gets from them is growing. You do not control the pace of the technology. You do control how quickly your company learns to use it.
Erik Brynjolfsson, who directs Stanford University’s Digital Economy Lab, calls rate of change a crucial concept. In a May 2026 interview he said that nobody knows how quickly a society can absorb a change of this size, and that there is already a large gap between what the tools can do and our ability to produce value with them. His conclusion is that we need to invest in understanding the transition as well as in the tools.¹
One test shows how quickly the capability is moving. OpenAI, the company behind ChatGPT, had experienced professionals set difficult tasks from their own work, and had other experts judge the results without knowing which came from a person and which from AI. In September 2025, the best AI model matched or beat the professionals nearly half the time. By March 2026, it was 83 percent of the time.² This is OpenAI’s own test, so treat the numbers with some caution, but the direction is clear.
Use inside companies looks very different. Brynjolfsson and his colleague Andrew McAfee report that in the companies they examined, roughly 3 to 5 percent of users achieved substantial gains, while most people used AI for modest tasks such as drafting emails and summarizing documents.³ Wharton professor Ethan Mollick is more direct in his observation. He says people are not giving AI hard enough tasks.⁴
Brynjolfsson’s recommendations are aimed at society, but three of them translate well to a single company. First, get better information. Find out where AI is changing the work in your business today, who uses it and with what results. Second, make it easy for people to move. As tasks change, people need training and a path into different work. Third, use AI to create new value. Some businesses focus on replacing employees, and Brynjolfsson argues the greater gain comes from creating things customers value.¹
Think of the last significant change in your business, such as a new system or process. How long did it take to become the normal way of working? If the answer is a year or more, and the tools improve every few months, the gap will keep widening unless you invest in closing it.
The pace of technology is dictated to you, but your rate of absorption is set inside the business. Be aggressive as you consider what type of work AI can take on and be intentional about helping your people adapt to change.
Sources
¹ Erik Brynjolfsson, interviewed by Nicholas Thompson, “What AI Is Really Doing to Jobs Right Now,” The Atlantic, video, May 27, 2026.
² OpenAI’s GDPval test. The September 2025 result is reported in The Decoder, “OpenAI says top AI models are reaching expert territory on real-world knowledge work,” September 26, 2025. The March 2026 result is from OpenAI, “Introducing GPT-5.4,” March 5, 2026.
³ Erik Brynjolfsson, Andrew McAfee and James Milin, “The ROI Gap: What 2026 Data Reveals About Enterprise AI Returns,” video, August 18, 2026. The three are co-founders of Workhelix, an AI consulting firm.
⁴ Ethan Mollick, interviewed by Simon Sinek, “The AI Skills Nobody is Teaching (And Everyone Needs),” A Bit of Optimism, video, June 16, 2026.