From Skills Data to Skills Experience: Why Most Skills Strategies Stall (And What Actually Works)

Most Skills Strategies Stop at the Spreadsheet

According to TestGorilla’s 2025 State of Skills-Based Hiring report, 85% of employers now say they use skills-based approaches, up from 73% just two years earlier. But only 38% maintain an enterprise-wide skills library (Mercer Skills Snapshot), up from 30% in 2023. Adoption is growing, but activation hasn’t kept pace.

In practice, most skills taxonomies sit in a system somewhere, reviewed quarterly at best, used for actual decision-making almost never. HR leaders know this. They just don’t talk about it at conferences.

“Most skills initiatives stall because they stop at representation,” says Nena Dimovska, Head of People Success at Semos Cloud. “The skills get mapped, they get validated, they get stored, and then nothing really changes.”

The distance between having skills data and doing something useful with it is the central challenge facing every skills based organization in 2026. Most organizations reduced skills to a data project, a classification exercise. Skills transformation is an experience design challenge. The question worth asking: do those capabilities shape how someone gets feedback on a Tuesday afternoon? Do they change how a manager prepares for a one-on-one? If the answer is no, the taxonomy is furniture.

This article draws from a recent Talent Alchemy podcast conversation between Nena Dimovska and Rinky Karthik, Senior Director of Product Success at SAP SuccessFactors. Their discussion surfaced a thesis worth unpacking: the jump from skills data to skills experience is where organizational transformation happens.

Skills Data vs. Skills Experience

Skills data describes what people can do. A skills taxonomy catalogs those capabilities across an organization, maps them to roles, validates them through assessments or manager input. This is foundational work, and it has real value. Every skills based organization needs it. The problem is that too many stop here.

Skills experience is what happens when those capabilities actively shape how people grow, get recognized, receive feedback, and find new opportunities inside the company. Every day. In the flow of work.

The difference shows up in who feels it:

Nena frames it simply: “The change happens when skills are treated as decision inputs, not just as some descriptors.”

Research from Deloitte shows that skills-based organizations are 63% more likely to achieve results. The fine print matters, though. The results lift comes when skills actually drive decisions, when managers use skill signals to assign work, when learning recommendations connect to observed weaknesses, when recognition ties to demonstrated capabilities. A taxonomy sitting inside an HCM does none of those things on its own.

That’s the line between companies that have invested in skills and companies where that investment changes behavior.

Why Skills Taxonomies Alone Don’t Transform Organizations

Fifty-five percent of companies now map skills to jobs, according to Mercer. That sounds like progress until you watch what happens after the mapping is done. In most organizations, the skills taxonomy becomes an expensive documentation project. It exists. It gets updated. Managers ignore it.

Nena has seen this pattern play out repeatedly: “You have a company that has a very strong skills taxonomy, but managers keep relying on their intuition when assigning work, when dealing with difficult projects, or just giving feedback on a performance review.”

The intuition part is important. Skills taxonomies fail to transform organizations because they don’t change the moment of decision. A manager preparing for a performance review doesn’t open the skills taxonomy. They open a blank document and write from memory. The data exists in one system. The judgment happens in another. There’s no bridge.

There’s also an evaluation failure. Organizations choose skills technology without first asking “what problem do we solve? what friction do we remove? what decision should become easier?” They start with the system and work backward, when the sequence should be reversed.

Nena’s prediction cuts deeper: “Skills data will become commoditized. Everybody will have these taxonomies. What will truly differentiate companies is how work feels.”

And 46% of executives already cite legacy mindsets, not technology, as the top obstacle to skills transformation (Deloitte). The hard part was always changing behavior. Taxonomies just gave us something to build while avoiding that conversation.

When AI Does the Homework, Managers Make Better Calls

Skills data only transforms an organization if it reaches the person making the decision, and in most companies it doesn’t. The manager preparing for a hard conversation and the taxonomy sitting inside the HCM exist in parallel worlds.

Nena pointed to a research direction that frames this well: “AI should do the homework while people remain accountable for judgment, for those high-stakes decisions.” She referenced an AI model covered by BBC Science Focus, trained on more than 10 million real human choices from psychology experiments. The finding: AI excels at pattern recognition and synthesis across massive datasets. People excel at contextual judgment, the kind that accounts for politics, trust, timing, and the things that don’t show up in any system. The combination outperforms either working alone.

(Our CEO Filip Misovski explored this human-plus-AI dynamic in depth in his piece on the Centaur Manager, applying it to leadership operating models. The principle is the same: augmented judgment beats pure automation and pure intuition.)

Take the manager one-on-one, the most common leadership moment in any organization. Most managers prep by opening a blank page and writing from memory. Maybe they glance at old notes. The conversation reflects what they happened to remember, which skews toward recent events and personal biases. Anyone who’s managed more than five people knows this.

Now add a skills intelligence layer. Agents continuously synthesize signals from feedback, recognition events, team meetings, and project outcomes. The manager walks into their one-on-one with a prep summary: skills the employee demonstrated recently, patterns over the quarter, recognition received, emerging growth areas. Same conversation. Same manager making the call. Better information feeding the judgment.

“The decision was evidence-based,” Nena says. “And this is where the magic happens.”

BCG reports that 72% of employees now use generative AI, but only 36% feel adequately trained. Gartner expects that within two years, managers will be leading both humans and AI agents. Most organizations haven’t prepared for that dual responsibility. Manager enablement is becoming an operational requirement, and skills intelligence is one of the inputs that makes it work.

Semos Cloud is building toward this with Manager Agents, currently in Enterprise Pilot, that translate work signals into timely recognition, targeted learning recommendations, and objective feedback. The one-on-one prep example comes from how the team at Semos Cloud uses the system internally. “What we’ve done even in our own team,” as Nena put it, gives a preview of what becomes possible when agentic AI meets daily manager workflows.

Culture as the Control System

Nena made one statement twice during the conversation, and the repetition was deliberate: “When AI takes on execution, the culture in the company will become our control system.”

McKinsey projects that up to 30% of current work hours could be automated by 2030. As AI handles more execution, what falls to humans is judgment and orchestration. Culture is what keeps those aligned. Or misaligned, depending on the organization.

“A high-performance culture doesn’t emerge from a taxonomy or a slogan or your vision board,” Nena says. “They’re established when leaders repeatedly reinforce the right behaviors over time.”

That word “repeatedly” carries weight. Culture isn’t built through a single offsite or a values rollout in January. It’s sustained through daily reinforcement. Recognition when someone demonstrates a valued capability. Feedback that connects performance to growth instead of checking a box. The kind of rewards that actually signal what the business values, not what fits a template. Those mechanisms do the work that slogans can’t.

For skills-based organizations, the implication is direct. Skills intelligence can tell you what capabilities exist across your workforce, where shortfalls are emerging, where strengths are concentrated. But that information only matters if the culture converts it into action. If capabilities go unrecognized, if development stalls in approval queues, if people with the right skills never get matched to the right opportunities, then the intelligence is just reporting on a problem nobody’s fixing.

HR’s role in this equation is evolving. As Nena puts it: “You can be this supportive function or you can become the architect of accountability.”

By 2027, Deloitte projects that 50% of enterprises using generative AI will have deployed agentic AI. That means AI agents making semi-autonomous decisions, surfacing recommendations, initiating actions. The governance question isn’t theoretical anymore. Who decides what the agent should optimize for? What behaviors should be reinforced? What should be flagged? Those are culture questions. And the skills experience layer is where they become operational.

The Internal Mobility Connection

Seventy-four percent of millennials and Gen Z workers say they’re likely to quit within a year if they don’t see development opportunities (Workplace Intelligence/Amazon).

Rinky Karthik raised this during the podcast: “A lot of employees leave the organization because they don’t know that there is a career they can grow into. There is no transparency.”

The compound problem looks like this. Gen Z wants visible growth. Growth requires internal mobility. Internal mobility requires living skills intelligence, real-time capability data that updates from actual work, not annual self-assessments completed reluctantly in December. That skills intelligence must surface to managers through AI so they can coach toward realistic opportunities. Managers then have evidence-based conversations about what’s possible.

Each link in that chain depends on the one before it. Break any one of them and the whole thing collapses back to “post the job internally and hope someone applies.”

“Previously it wasn’t easy to do it at scale,” Nena notes. “It was doable for small organizations, but at scale it was very complex. Now with all this technology at hand, it’s really possible.”

The technology she’s referring to isn’t a single product. It’s the convergence of skills taxonomy infrastructure, AI-powered signal detection, and agentic AI that can surface the right information to the right person at the right moment. A skills based organization that actually retains talent needs all three layers working together. The taxonomy provides the vocabulary. AI provides the intelligence. The skills experience makes it real for the individual employee who’s deciding whether to stay or update their resume.

Building a Skills Experience Layer

A skills experience layer connects your HCM system to your people, turning raw skills data into moments that change behavior.

Nena’s starting point is blunt: “Begin by defining the human problem.” What decisions are managers struggling with? Where does cognitive load pile up? Where do employees feel invisible? Start with those friction points, not with system architecture or vendor selection.

The operating model argument matters here too. Skills scale when they’re treated as a cross-functional capability, as part of how the company operates. When they live exclusively inside HR, they get championed for two quarters and quietly deprioritized. Every stakeholder needs to see the same evidence and pull the same levers, or, as Nena puts it, “skills will be just a debate, not a mechanism.”

A skills experience layer has three functional components, though they look different in every organization:

Sensing is where systems detect skills from real work signals. Recognition events reveal what capabilities people actually demonstrate on the job. Feedback data captures how those capabilities develop over weeks and months. Even meeting summaries can surface coaching moments that would otherwise disappear. The shift is from periodic self-reporting (“fill out your skills profile in December”) to continuous signal capture from the work itself.

Deciding is the manager layer. Before a one-on-one or a work assignment, skills intelligence informs the judgment call. The manager still owns the decision. What changes is the quality of information available, and that change compounds over dozens of conversations per quarter.

The third layer, experiencing, is what the employee feels. Recognition that arrives tied to a specific skill demonstrated last week. Learning recommendations connected to a skill they’re working to build. Growth conversations with context instead of platitudes. This is where the skills based organization stops being an HR label and starts showing up in someone’s Wednesday afternoon.

Nena captures the design principle: “Personalization is not analytics, is not dashboards. It’s timing, relevance, and shared purpose.”

Where does this live technically? Inside the existing HCM ecosystem. SAP SuccessFactors, Workday, and Oracle provide the backbone, the system of record, the infrastructure. Complementary AI layers activate the intelligence sitting inside those systems. Semos Cloud operates in this space as an experience layer within the HCM, a bridge between the data that exists and the moments where that data should influence a human decision.

The skills intelligence platform market was valued at $2.37 billion in 2024 and is projected to reach $16.7 billion by 2033, according to Growth Market Reports. That 7x growth over nine years tells you where enterprise budgets are heading.

Where This Leaves Most Organizations

Every large enterprise will have a skills taxonomy within the next few years. Nena’s prediction on that is probably right. The taxonomy becomes table stakes.

So then what separates a functioning skills based organization from one that just claims the label? The difference shows up at the experience layer: does that data change how a manager prepares for a hard conversation? Does it help an employee in Bangalore see a realistic path to a role in Singapore? Does recognition connect to demonstrated capability or just tenure milestones?

“The future of work is about creating environments where AI does the heavy lifting, but people own the call, and culture is the control system,” Nena says. That’s the clearest framing of where this is heading.

The organizations building skills based organization infrastructure around experience will have a two to three year head start by the time their peers finish selecting a taxonomy vendor. Given the pace of 2025 and the first months of 2026, that window won’t stay open long.

This article is based on the Talent Alchemy podcast conversation between Nena Dimovska (Head of People Success, Semos Cloud) and Rinky Karthik (Senior Director of Product Success, SAP SuccessFactors).

Related posts

Analyst Spotlight: AI Readiness in HR – From Hype to Human Value  

AI capability is accelerating. Organizational change is not. As expectations rise, HR must bridge the growing gap between innovation and adoption. read more

Introducing Celebrate Inside Workday: A Built on Workday App for Employee Milestone Celebrations

read more

Why 2026 is The Year Employee Communications Becomes a Core System 

Employee communications has evolved into a core enterprise system. In a hybrid and deskless world, organizations need a reliable way to reach every employee, reduce noise, and drive alignment at scale. read more