In September, we re-ran our AI Proficiency Survey to over 5,000 knowledge workers across the US, UK, and Canada. The full report on our findings will launch this month but, to whet your appetite, here are the top 3 findings that surprised us – and that we think should motivate you to take another look at your 2025 AI strategy.
As a reminder, our AI Proficiency Survey asks respondents questions about their AI usage and then tests their understanding of responsible AI use and ability to prompt. We give each respondent a score out of 100 based on their AI usage, knowledge, and prompting abilities, and then bucket users into one of five categories based on their scores:
- AI Expert – 1% of the workforce
- AI Practitioner – 8% of the workforce
- AI Experimenter – 34% of the workforce
- AI Novice – 47% of the workforce
- AI Skeptic – 11% of the workforce
We also asked respondents questions about their AI attitudes and behaviors including whether they’ve taken AI training, how their company and direct manager talk about AI, and where they sit in the organization (function, level, etc.).
Here are the top three findings that tell us most orgs aren’t ready to deploy AI.
1. AI training is largely ineffective
Pretty convenient coming from an AI training company, we know, but the numbers don’t lie. At first glance, AI proficiency does seem to scale with access to training – only 4% of AI Skeptics and 23% of AI Novices report receiving training on AI vs. a third of Experimenters and Practitioners and nearly half of AI Experts.
But when we looked closer at respondents’ scores on AI usage, knowledge, and prompting, the only thing training appears to meaningfully influence is how frequently someone uses AI.
It doesn’t increase AI knowledge or prompting in any meaningful way.
This tells us that most AI training today isn’t driving proficiency – it’s likely centered around company policies rather than tactical, hands-on teaching about AI’s capabilities, risks, and how to get good output from it.
2. Direct managers can roadblock (or boost) AI adoption
How direct managers talk about AI (or don’t) can make or break their team’s adoption and proficiency, even if the overall company policy and rhetoric promotes AI use.
Among companies that explicitly approve of AI use, 14% of employees whose manager also explicitly encourages or expects AI use are an Expert or Practitioner. But that number is cut in half when the manager disapproves of AI use – and AI Skeptics grow by 3x.
The same is true in companies that are silent on AI use. If the direct manager encourages AI use, employees are more than twice as AI proficient as those in departments where the direct manager discourages AI use.
You need your front-line managers on board, because they can take you off message and off strategy.
3. Senior employees are more proficient and less skeptical of AI than their more junior counterparts
Senior employees (director and above) are significantly more AI proficient than more junior individual contributors. 12-13% of senior employees scored as AI Experts or Practitioners vs. only 5% of individual contributors (ICs).
ICs are also 3x more skeptical about AI – with 16% of junior employees scoring as AI Skeptics vs. just 5% of senior employees.
This constitutes potentially massive wasted potential, considering ICs are more likely to own the repetitive and manual tasks that AI is so good at automating. If these companies were to graduate the 50% of junior employees that rank as Novices to Experimenter level proficiency, their ROI is likely to double.
How you can get ahead of these trends for 2025
Our conclusion, based on the total findings, is that most companies are actually not ready to deploy AI to their teams – who are not in a position to use AI safely or effectively. Here’s how you can turn the tide for your team:
- Run this AI Proficiency Diagnostic internally. Measure your team’s AI usage, knowledge, and skill and we’ll compare it to the broader benchmark and your industry to provide personalized recommendations for your AI strategy.
- Attend our AI:ROI Conference. Learn from the successes and challenges of real heads of AI from leading companies like Moderna and S&P Global and integrate them into your 2025 planning.
- Roll out hands-on training. AI training needs to be hands-on-keyboards in order for it to work. Get your team unlimited access to our AI Academy where they can access live and on-demand versions of our foundational and functional AI courses that are built around immediately applicable outcomes.
