For years, organizations struggled to truly see the skills within their workforce.
Now, Microsoft is taking a bold step forward — using AI to infer skills dynamicallybased on how employees engage with Microsoft 365 apps like Outlook, Teams, Word, and Copilot itself.
What Skills Agent can do:
➔ Build skills profiles automatically — based on real work, not just resumes or self-assessments.
➔ Map skills across teams and business units — giving leaders a live view of workforce capabilities and gaps.
➔ Highlight skill adjacencies — identifying not just current skills but future potential based on behavior patterns.
➔ Support internal mobility and growth — suggesting career paths personalized to employees' real strengths.
➔ Connect learning to real needs — linking skill gaps directly to upskilling opportunities.
Why this matters:
— It gives organizations a dynamic, data-driven view of their evolving workforce.
— It moves companies closer to operating in a truly skills-based way, rather than relying on static roles or outdated titles.
Here are the current limitations I see:
➔ Skill inference is not skill assessment.
Observing usage within Microsoft apps shows what tools people interact with — but doesn’t reveal the depth of mastery.
Someone might spend all day coding in a particular language — but are they an expert, still struggling to master it, or just using it inefficiently? Inference alone can’t fully tell that story.
➔ Focus appears mainly on technical and digital skills.
It seems clear this solution is optimized for mapping technical capabilities. Yet, as AI increasingly automates technical tasks, soft skills — like leadership, adaptability, and collaboration — are becoming even more crucial for success. These human skills are much harder to infer from system usage data.
➔ It depends heavily on Microsoft ecosystem usage.
If employees work extensively in other environments — Slack, Google Workspace, GitHub, Notion, and more — key parts of their skillset may be invisible to Skills Agent.
Our take:
This is an incredible innovation — and a major leap for workforce intelligence.
But it’s one piece of the broader puzzle.
Organizations aiming to truly unlock human potential will still need to combine skills inference, validated skills assessments, and thoughtful evaluation of the uniquely human capabilities that AI cannot easily measure.
Pictures source: Microsoft
Links to more reading on the topic:
Microsoft Launches People Skills In Copilot, Altering The HR Tech Market
