Talent leaders have spent the past few years focusing on how to assess durable skills in the hiring process. These days, AI can write a resume, tailor a cover letter, and coach a candidate through a behavioral interview. Signals that once took effort to produce, and told employers something real about a candidate, now take a prompt. So companies are adding live exercises, whiteboard sessions, and follow-up questions, all trying to see a skill in action before they make an offer.
That work matters, and it has limits. Hiring can be a challenging place to get a good read on durable skills like judgment, communication, and collaboration. When hiring, an employer is trying to predict how a stranger will apply those skills on the job, based on a few hours of contact.
Inside the company, though, the picture is different. Employees have been demonstrating their skills for months or years. Their managers have watched them handle a difficult client, rescue a stalled project, and bring a skeptical teammate along. That’s the strongest kind of evidence there is: real work, under real conditions, over time. And no hiring assessment can fully replicate that.
The Evidence Is Already on the Payroll
In conversations with talent leaders at large employers this year, as part of our work with Business Roundtable, it was clear that the problem is rarely a shortage of information. Managers know who has good judgment. Peers know who makes a team better. What’s missing is a reliable way to capture that knowledge so others can use and compare it. A manager’s sense of an employee’s leadership lives in that manager’s head, measured against a standard only they can see. One talent leader told us that what they call good leadership could be very different from what a colleague calls good leadership. Multiply that across hundreds of managers and the evidence can’t be compared and used effectively, even though it exists.
When evidence for durable skills exists within a company but can’t be compared, internal talent decisions can rest on less standardized methods of assessing that evidence than external hiring decisions do. A company might put an outside candidate through a structured interview with a scoring rubric and trained evaluators, then fill the next role internally because someone’s manager spoke up for them. The stranger got a rigorous assessment. The person the company knew best got a recommendation.
When companies address this, the results can be substantial. One company in our research, profiled in our recent report with Business Roundtable, shifted its focus toward demonstrated skill and potential. Three times as many of its roles are now filled by internal candidates as three years ago. In this case, the company didn’t find new, outside talent. Instead, it started looking at what its people had already done.
Using What’s Already There
Capturing internal evidence doesn’t require a new platform. It requires companies to apply the discipline of good hiring to their own people. Our report lays out four properties of a strong skill signal, and with small adjustments they fit internal decisions as well as external ones.
- Behavior-anchored. Internal evidence is made of behavior. Employers don’t need a simulated case when they have a real one. The question is whether anyone captures what the person did and how it moved results for the business, rather than how they came across in a meeting.
- Standardized. This is where most internal processes break down. If managers define strong performance differently, their assessments can’t be compared. Clear descriptions of what strong, adequate, and weak performance look like on a given skill go a long way toward closing that gap.
- Multi-source. Inside a company, sources are everywhere. Peers, project leads, internal clients, and managers have all seen the work. Capturing and synthesizing these sources in a consistent way generates a far fuller picture than any single manager’s judgment.
- AI-resilient. Here internal evidence has a built-in edge. Years of observed work provide evidence no chatbot can produce.
Promotion Is Only Half of It
The same blind spot shows up in development, where the cost is harder to see. Promotion asks whether someone is ready. Development asks where a specific gap is and whether it’s closing, which means measuring the same skill repeatedly against the same standard.
Companies do this well for technical skills; they track certifications and required training, and they define career paths in detail. Durable skills are often relegated to a leadership program for a selected few. Everyone else is left to build communication and problem-solving on their own, with no one measuring whether they’re improving.
As AI takes on more technical work, durable skills increasingly separate strong performers from adequate ones. A company that can see those skills clearly in its own workforce will develop them faster and put people in the right roles more often than one that reserves its most rigorous assessment for outside candidates.
For any talent leader, a useful question to start with is this: what do we already know about our people that never makes it into a decision? At many companies, the answer is a lot.
