Over the past ten days, 17 people completed real assessments of their own roles on automatable.me. Six of them were Product Managers—the most-assessed profession in this small window. And here's what jumped out: their average effective automation exposure landed at exactly 28%, matching the platform-wide average to the decimal point.
That's worth pausing on. Product Manager work is almost entirely digital input and output—roadmaps, specs, feedback loops, dashboards. On a naive reading, these roles should light up as highly automatable. But they don't. Instead, the real exposure metric—which weights in the tacit knowledge required to do the work well and the trust-based relationships embedded in the role—pulls the number down substantially. PMs describe tasks that look procedural but depend heavily on judgment about tradeoffs, team dynamics, and product intuition accumulated over years. Removing the person doesn't just remove a process; it removes a decision-maker whom stakeholders trust.
A caveat: this is still very thin data. We have 17 real assessments this period and 361 total on the platform, much of it seeded with synthetic profiles. We're not confident yet. But the early signal is interesting: the roles that *feel* most automatable—heavy on digital process, light on physical presence—aren't necessarily the ones with the highest real exposure when you account for what the work actually requires.
If you're in a role that's mostly digital—PM, analyst, designer, engineer—the question worth asking yourself isn't whether your tasks can be digitized. They already are. The question is: what accumulated judgment or trust are you providing that a system alone can't?