Most AI Skill Courses Miss What’s Actually Needed: AISA Report
Per the AISA study, average AI fluency score is 46.1, while 69.3% score below Proficient, showing gaps in corporate AI training priorities.
Most corporate AI training programmes are designed before organisations assess where employees’ skills fall short. New data from AI fluency assessment platform AISA suggests this approach may be directing training towards areas where employees are already relatively strong, while overlooking weaker skills.
The latest AISA assessment report found a clear AI skills gap across roles and industries. The average AI fluency score was 46.1 out of 100, placing participants in the Developing tier. A further 69.3% scored below Proficient.
However, when AISA asked 89 people which single AI skill they would prioritise improving across their teams, their choices did not reflect the gaps identified in the assessments.
More than half (51.7%) selected building AI into workflows. This was already the highest-scoring dimension in AISA’s assessments. In contrast, only 4.5% said they would prioritise verifying AI output, despite this being one of the weaker areas measured.
Technical understanding of how AI models work was the lowest-scoring dimension, with an average score of 39.0 out of 100. AISA said this area is rarely covered in corporate training.
External research points to a similar gap. Docebo’s 2026 AI Readiness Gap report found that 85% of employees feel their AI training does not help them use AI effectively in their roles.
AISA recommends a four-step approach to corporate AI training: measure, target, train and re-measure. Organisations should first establish a baseline to identify the weakest skills within each team. Training can then be directed towards those gaps, followed by another assessment after 8 to 12 weeks to measure progress.
According to AISA, 98% of people who retake its assessment land in the same or an adjacent tier. The company says this means changes between assessments reflect real skill movement.
“Managers ask for the skill they can see. Someone automating a report is visible on a Tuesday. Someone who never checks the numbers AI gives them stays invisible until the day it goes wrong. I spent years in HR and training rooms, and the pattern is the same every time: if you don’t measure first, you end up polishing what’s already decent. AI fluency is the new currency, and you can’t spend what you haven’t counted,” said Ozan Dagdeviren, Founder and CEO of AISA.