crossed 1,000 AI skill assessments — the job role breakdown surprised us

hit a milestone this week on aisa.to — over 1,000 people have now taken our conversational AI skills assessment. finally enough data to break it down by job role.

the finding that caught me off guard: product managers scored highest overall. not engineers, not data scientists. PMs. my theory is they're already in the habit of breaking problems down and evaluating output critically, which maps well to how you actually use AI effectively.

marketing had the worst safety scores. not because they're careless — they're under pressure to ship content fast and verification feels like friction. 36% of everyone we assessed had no functional safety practice at all.

full breakdown with numbers: https://aisa.to/blog/ai-skills-by-job-role-2026-data

from a product perspective this data is shaping the next version. we're considering role-specific assessment paths so the conversation adapts to what actually matters for a PM vs an engineer vs someone in ops.

curious if other founders are seeing similar patterns. does the role someone's in predict how well they use AI tools?