Knowledge workers are using generative AI to write emails, summarize information, generate content, draft technical copy, brainstorm ideas and analyze data. Yet less than a third of companies that employ such workers have a formal AI strategy in place. And “dangerous divides” separate knowledge workers from the leaders to whom they ultimately report.
The figures and the alert are from a survey conducted by the work management platform supplier Asana and the AI safety and research startup Anthropic. Researchers from the two received completed responses from around 5,000 knowledge workers in the U.S. and U.K.
Noting that more than half of knowledge workers now use gen AI, the report authors observe that organizations move through five stages along the path to full AI integration—skepticism, activation, experimentation, scaling and maturity. Along the way, they suggest, AI-primed operations do well to take five questioning strides—all starting with a “C”—to successfully traverse the five stages.
1. AI Comprehension: How well do your employees understand how to use AI?
To reach AI maturity, employee understanding of the technology is crucial, and it varies significantly across stages, the authors write. At stage 1, “AI skepticism,” only 2% are familiar with generative AI basics, 6% have a strong understanding of its capabilities for their work, and 20% use AI weekly at work. More:
By stage 5, “AI maturity,” familiarity with generative AI surges to 35%, 53% have a strong grasp of its capabilities, and 93% engage with AI each week.
2. AI Concerns: What issues are top-of-mind for employees regarding AI?
As workers develop basic AI proficiency in stages 2 and 3—“AI activation” and “AI experimentation”—new fears emerge, Asana and Anthropic point out. Contributors become increasingly concerned about others’ perceptions of their new AI use, worrying that relying on AI might be seen as taking shortcuts or producing inauthentic work. More:
Data shows 29% of workers worry about being perceived as lazy and 25% feel like frauds for relying on AI to complete tasks.
3. AI Collaboration: How do employees work together with AI?
Workers who see AI as a teammate are 33% more likely to report productivity gains from using AI at work, compared to those who consider it a tool. The authors further note that, as workers use AI more frequently, they begin to recognize its potential for collaboration and its capacity to assume more complex roles within their workflows. More:
Workers who interact with AI on a daily basis are significantly more likely to prefer AI to act like a teammate (22%) compared to those who use it monthly (16%).
4. AI Context: What AI policies, guidelines and principles frame the organization’s outlook on AI?
At stage 1 (AI skepticism), only 2% of employees report that their organizations have defined AI principles, compared to 34% at stage 5 (AI maturity). By stage 5, organizations not only recognize that well-defined AI policies and principles are crucial for regulatory compliance, but they also view them as strategic assets that can differentiate them in the market.
These policies and principles provide a clear framework for employees to follow when using AI, ensuring ethical, consistent and responsible usage across the organization.
5. AI Calibration: How is AI effectiveness and value measured in your organization?
To effectively calibrate AI, organizations must engage their workforce in the evaluation process, the authors state. At stage 1 (AI skepticism), only 17% of workers say their organizations actively collect employee feedback on AI tools, which may contribute to the modest productivity improvements observed at this stage.
By contrast, at stage 5 (AI maturity), this practice becomes well-established, with 91% of workers reporting that their organizations actively incorporate employee feedback into the AI calibration process.
Download the full report.
