Who validates the AI's suggestions, the employee or their manager?

The employee, by default. Each person gets their inferred skill set pushed to them through a skill assistant that sits in Microsoft Teams, and works through an accept, reject or add flow, setting or adjusting proficiency along the way. We can also route it through Slack or a bot you already run. Their manager sees what the employee decided, as an FYI.

That default is deliberate. When a manager disagrees with a skill an employee has claimed or dropped, we want that to become a growth conversation between the two of them, not a silent override in a system.

Some organisations want the stricter version, where skills route to the manager first and only count as validated once the manager has signed off. We support that as well. Which route you take is a governance decision you make when you design the rollout, and it is worth settling early, because it changes what you tell employees the exercise is for.

Proficiency is handled separately. Each skill carries three values side by side: the employee’s self-assessment, the manager’s assessment, and the level inferred by our AI or by a connected third-party source. Where what someone claims and what the data shows disagree, that stays visible on the profile.

Job profiles follow a different path. Nominated subject-matter experts review the AI-generated job-to-skill profiles in the TechWolf console, a separate surface from the employee-facing assistant, and changes to the skills taxonomy go through their own approve-or-reject step before anything is written back into Workday or SuccessFactors.

So the thing to settle before launch is which of the two routes fits your organisation. More on how we handle skill governance.