Microsoft People Skills: What it sees and what it misses

Earlier this year Microsoft announced its new skills solution, tightly integrated into Microsoft 365 called People Skills. The new technology promises to map your organisation’s talent using AI, but its vision is limited by what it cannot see.

The technology takes a novel approach to skills management:

1. You start by enabling their built-in skills taxonomy or importing your own.
2. AI periodically scans all activity within the Microsoft 365 suite of products to assign skills to users (this process is known as “inferencing”).
3. Confirmed skills are rolled up to various dashboards and views.

The inferencing is the important part of the technology here. It passively observes activity within the Microsoft suite of tools and translates them into implied skills. It’s the difference between what you’ve done (“wrote a deck on marketing strategy”) and what you’re capable of (“strategic planning”). However, its usefulness depends on what signals are available and the accuracy of inferencing.

Microsoft says People Skills helps you “understand the backbone of your company – your people”. Although as we discover, the reality is closer to understanding one or two vertebrae in that backbone.

Why is a skills framework important?

When evaluating the usefulness of People Skills, we first need to understand the problem it’s solving. Increasingly organisations are adopting or developing their own skills frameworks. There are two key pieces to an effective organisational skills framework:

1. Defining a skills taxonomy covering the necessary skills required by roles in an organisation, and how they relate to each other.
2. Undertaking skills verification to assess an individual’s proficiency across each skill (as it applies to their role, or related roles).

Doing both well unlocks benefits essential for workforce planning:

• Visibility over workforce mobility and where resources can be reallocated based on increased demand.
• Skills gaps in the existing workforce, including the ability to forecast and make more informed hiring decisions.
• Succession planning and more meaningful career conversations based on data instead of anecdotal evidence.
• Learning needs analysis to determine where training gaps exist and how to better utilise L&D budgets.

For these reasons, we recommend every mature business invest in developing a solid skills strategy to build resilience. The question is, how well does Microsoft People Skills rank on the above?

What is People Skills good at?

There’s no disputing that Microsoft got a few things right with People Skills. It comes with a plug and play skills taxonomy pulled from LinkedIn data, making it trivially easy for businesses without an existing framework to get started. We’re a big believer that organisations should tailor their skills taxonomy but having something out of the box is a massive advantage. And to MIcrosoft’s credit, they still allow you to bring your own taxonomy.

More importantly, People Skills gets out of the way of employees. It sits in the background, quietly assessing and assigning skills to individuals as they complete their day-to-day tasks. This “in the flow of work” skills verification is its greatest strength because it doesn’t detract from an employee doing their work. It also allows employees to confirm or reject the skill if AI gets it wrong.

This is perfect for knowledge workers who spend most of their working life in the Microsoft suites of tools. Unfortunately, it’s also People Skills’ biggest failing.

What are the biggest gaps in People Skills?

Knowledge Workers are a natural fit because they demonstrate skills inside the systems being monitored. A Financial Analyst will do most of their work in Excel, Powerpoint and across email – all systems that are monitored by People Skills for skills inference.

However, there are entire classes of workers who live outside the Microsoft ecosystem. Engineers perform most of their work in GitHub, JIRA or other task-management platforms. Marketers might be using HubSpot, SalesForce or other CRMs. Customer Support could be using Zendesk, ServiceNow and other ticketing systems.

All of these are largely invisible to how People Skills identifies skill demonstration. This means even within the world of knowledge workers, entire departments might be missing from the skills tree. To Microsoft’s credit, they’re building connectors to bring in more data from third-party systems so this current limitation that might be overcome in time.

Even further removed are frontline workers who do their work in the real-world, completely removed from the digital ecosystem. That’s the largest part of the workforce in industries such are healthcare, retail, manufacturing and many others. Right now People Skills doesn’t have a way to monitor their activity or determine their skills.

Even when work is undertaken within the Microsoft ecosystem, there are still a few issues to overcome.

People Skills doesn’t evaluate proficiency

We’re strong believers that skills should be evaluated against proficiency levels. There’s no industry consensus on what those levels should be – some frameworks use 7, others use 5, and each level is often labelled differently. Most experts agree that individuals vary in their expertise, from novice to expert.

Peoples Skills doesn’t account for this nuance, you either have the skill or you don’t. This limits the usefulness of skills verification, particularly when it relates to workforce mobility and succession planning.

Worse still, activity doesn’t mean you have a skill. Just because you mention Agile Coaching on Teams, or write a slide deck on the topic doesn’t mean you’re a competent Agile Coach. Without understanding exactly how the AI evaluates skills (this part is proprietary and also raises some AI governance questions), it’s hard to validate how accurate the skills verification piece is. Time will tell, but on a surface level it sounds susceptible to false positives.

This raises the question of whether you should enable People Skills at all.

Our recommendation

People Skills isn’t suitable as the basis of a skills framework, but it can help contribute to a wider skills strategy. It’s undeniable that skills verification is hard and any indicator is better than no indicator. If you don’t already have a skills taxonomy, this is a great place to start and a low-effort way to dip your feet into the world of skills.

If you’re looking to further develop an existing strategy then treat People Skills data as a weak signal to supplement additional sources of skills verification. Best practices should take a layered approach, including all of the following:

• Performance appraisals – manager and peer reviews remain the most context-rich measure of applied skills.
• Learning undertaken – verified completion of relevant courses, including assessment results.
• External qualifications and certifications – verified credentials from professional bodies or external training providers.
• Observational checklists – particularly valuable for frontline and hands-on roles where digital signals are limited.
• 360-degree feedback – insights from peers, managers, and direct reports to validate behavioural and leadership skills.
• Project and work portfolio evidence – real-world deliverables and outcomes that demonstrate capability. People Skills fits here.

There’s going to be fragmentation, that’s unavoidable. Skills data comes from multiple sources, in different formats with varying degrees of confidence. The goal is to build a composite picture that balances automation with human judgement.

Microsoft People Skills is effective because of its seamless integration into an existing ecosystem. Many analysts recognise this as a huge disruptor in the market due to its accessibility and scale. They’re not wrong, Microsoft has cleverly positioned the product and many will adopt it out of convenience. However, more mature organisations should be weary of its limitations and weigh up the trade-offs with a more comprehensive skills strategy.