L&D Doesn't Need a "Seat at the Table." It Needs Real-Time Capability Intelligence.

For a decade, the goal was the seat. Get L&D into the strategy conversation, and the rest would follow.

RedThread Research's 2026 State of L&D report suggests that goal is slipping away. L&D's participation in business strategy discussions has dropped for the third straight study. Read on its own, that sounds like a function losing relevance.

But the top 25% of L&D organizations in RedThread's data tell a different story. They aren't winning the seat back. They've stopped asking for it. They are embedded in the functions they serve, closer to where decisions get made, and they score at least twice as high on every dimension of learning culture. As RedThread's Dani Johnson put it, the best teams see themselves as "a strategic-tactical team on the ground" rather than a pillar that needs an invitation.

That shift changes what L&D needs. Not a better pitch for the boardroom. A better signal on the ground.

Employees already left the building

The most telling finding in RedThread's study isn't about L&D's influence. It's about where development actually happens. The methods employees rely on most, manager feedback and stretch assignments, are ones L&D doesn't own. For the first time in the study's history, none of the traditional L&D-owned methods ranked in the top 10.

Workera's 2026 State of Skills Intelligence Report, a survey of 1,000 U.S. enterprise employees, shows the same gap from the employee side:

• 58.3% say their organization offered AI-specific training in the past 12 months, up from under a quarter the year before.
• Yet 56.4% say the biggest barrier to building AI skills is that no time is allocated during work hours, and 42.5% cite a lack of relevant materials.
• 21.5% say their training is too basic or outdated, up from 16.8% in 2025. It's one of only two barriers that got worse this year.
• 84.3% spend five hours or less per week on development. 12.8% spend none.

More access to training. Same barriers. Worse relevance. The program got bigger and the gap didn't move, because the program was built around what L&D could deliver, not around what each person actually needed on a given Tuesday.

Completion metrics are a relic

If employees are developing through feedback and stretch work instead of courses, then course completion is measuring the wrong thing. It always was. A certificate proves someone consumed content. It says nothing about whether they can apply it.

The cost of that illusion is now well documented. Josh Bersin puts corporate learning spend at $400 billion, with 74% of companies still unable to keep pace with skill demand. IDC projects that more than 90% of global enterprises will face critical IT and AI skills shortages this year, and 48% of IT leaders already report that a lack of AI skills is holding back initiatives. And in Gartner's 2026 State of Skills research, published this July, 59% of CHROs say skill needs are changing faster than their organization can build plans to support them.

Completion data can't explain any of that, because it was never designed to. It tracks activity. The business needs capability. This is where most L&D leaders find themselves at budget review: a program they believe in, a stack of completion reports, and no way to prove to leadership that anyone got better at the job.

The people side is just as expensive. In the State of Skills Intelligence Report, 42.5% of employees say they've been passed over, held back, or mis-rated because someone misjudged what they could do. Nearly 60% say AI skills are weighted equally or not prioritized when their organization makes promotion and assignment decisions. When the only evidence of growth is a transcript, managers fall back on gut, and gut gets it wrong two times out of five.

What "embedded" actually requires

RedThread's high performers aren't embedded because they attend more meetings. They're embedded because they can answer the question a business leader actually asks: who on this team can do this work right now, and what will it take to get the rest there?

Answering that requires real-time capability intelligence. Two things have to be true.

First, the baseline has to be verified, not inferred. Resumes, self-ratings, and inferred skills profiles are perception. Workera's platform data shows that seven in 10 employees misjudge their own ability levels, overestimating or underestimating what they can do. Adaptive, granular assessment built on Evidence-Centered Design replaces that guesswork with measured proficiency at the skill level, so L&D can see the exact delta between where someone is and where a project needs them to be. That precision is what lets you bypass what people already know and point them only at the gap. It's also what makes a stretch assignment a calculated bet instead of a hopeful one.

Second, the signal has to keep moving after the baseline. A test taken in January doesn't tell you who's ready in June. Skills have a half-life now, and digital and AI skills expire faster than the rest. Workera measures proficiency continuously across the talent lifecycle, re-verifies as people grow, and flags skill drift before it shows up as a failed project. The output is a real-time skills velocity metric: not who completed what, but how fast the workforce is actually getting better, by person, role, and program.

Employees are more open to continuous measurement than most leaders assume. 41.1% say they would opt in if they owned the data, and another 31.2% are undecided. That leaves 72.3% who aren't opposed, and one condition that decides which way they go.

What this looks like on Monday morning

With a verified baseline and a living signal, L&D stops being a content function and starts operating the way RedThread's top quartile already does:

• Match people to immediate gaps, not generic paths. When a team needs someone who can orchestrate AI agents by Q1, you know who's closest today and what specifically stands between them and ready.
• Turn stretch assignments into a system. Pair demonstrated capability with the next-level work that builds it, and measure whether the stretch actually moved the skill.
• Give managers evidence to coach with. Only 26.6% of employees receive coaching from a manager or executive a couple of days a week or more; 21.5% never do. Capability data gives a 1:1 something concrete to work from.
• Report what moved, not what was watched. Skills velocity, proficiency growth, and readiness against business priorities are numbers a CHRO can take to the board. Completion rates are not.

The results compound quickly. Siemens Energy saw a 62% improvement in generative AI skills in two weeks once development was targeted at verified gaps rather than delivered as a catalog.

Stop asking for the seat

RedThread's data makes a case most L&D leaders already feel: the seat was never the point. Influence comes from being the function that knows, in real time and with evidence, what the workforce can do and how fast it's getting better. That is the proof upskilling has been missing. Show leadership what changed, not who finished, and the seat stops being something L&D has to ask for.

That's the gold standard in measuring human potential in the AI era, and it's available today.

Read the full 2026 State of Skills Intelligence Report to see what 1,000 enterprise employees said about training, coaching, and AI at work. Or request a demo to see a verified, real-time view of workforce capability in one platform, the kind of proof that shows leadership what changed.

‍