This Week's Term: Skills-Based Organization (SBO) — the structure of work, mobility, learning, and decisions organized around discrete skills and tasks rather than jobs and ladders.
The phrase has been popularized by Deloitte's Human Capital Trends research since 2023 and is now being operationalized by enterprises and HR tech vendors.
The clearest analogy is Netflix. Most internal job systems are like a video store with shelves: Drama, Comedy, HR, Sales, Engineering. You walk in, find the section, pick something. A skills-based organization is what happens when the shelves dissolve. The system knows what every employee can do, what every job requires, and what learning paths bridge gaps between the two. It surfaces opportunities the way Netflix surfaces shows. TechWolf's Mikaël Wornoo calls this "personalizing talent management processes at scale."
The three components of a working SBO
A working SBO has three components.
A skills ontology, the structured map of the skills in your domain and how they relate.
A skills graph for every employee, an inferred, continuously enriched view of what each person can actually do.
A talent marketplace, the system that matches the two and surfaces opportunities.
Vendors infer the employee graph from the digital footprint people already leave: previous roles, learning, performance reviews, internal communications. Validation by employees and managers gets accuracy into the 85 to 95 percent range. The Findem case study at RecruitMilitary in this issue's tools section shows what this looks like in practice on the recruiting side.
Why SBO matters in 2026 (it's not just about humans)
There is a second reason SBO matters in 2026 that the original Deloitte framing did not anticipate. Skills are now also the foundation for AI agents.
The Skills format I covered in the Writing Skills post from Issue #26, and the broader cross-platform Skill ecosystem spanning Claude, OpenAI, Cursor, Gemini, and others, treats a skill as a packaged unit of know-how an agent can load on demand. The same abstraction that lets a human contribute without a fixed job description lets an agent contribute the right capability at the right moment.
And because skills are transferable, you can teach your AI tools and agents to do things your way at three levels at once. An individual leveling up their personal AI assistant. A team encoding how its work actually gets done. An organization codifying its operating model so any agent that joins can pick it up. SBO is no longer just about people. It is about how the whole hybrid workforce, human and agent, learns and improves.
Why this connects to the dual mandate
This is the bridge to the HR dual mandate framework I write about in this issue. Mandate 1, transforming HR itself, needs an SBO underneath to deliver. Mandate 2, redesigning the organization around hybrid human-AI teams, needs an SBO that treats agents as workforce members with their own skill profiles. Both mandates fail without it.
Practical implication for leaders
Ask two questions of the AI rollout you are sponsoring.
- Is it sitting on top of a real skills graph, or a wishlist?
- Is the skills work you are funding scoped only to HR, or built so the same skills guide how your AI tools and agents operate?
If either question lands on the wrong side, the rollout will fail in interesting ways. Workday, SAP, and Oracle now ship SBO functionality natively in their core HR systems. For most organizations, the first move is not to buy a new platform. It is to turn on what they already own, pair it with a data layer, and treat the skills work as a shared investment.
Skills work pays twice in 2026. Once for the people you employ. Once for the agents working alongside them.
For deeper learning
For a focused walkthrough of how a working SBO actually delivers, including the realism about data quality and examples of measurable outcomes across his customer base, listen to Mikaël Wornoo's appearance on the Future Ready Leadership podcast.
Your action step
This week, ask your HR or HRIS lead one question: do we have a skills ontology and a per-employee skills graph today, or are we still relying on job titles and self-reported expertise? If the answer is the latter, your AI strategy has a structural data problem, not a tooling problem. Fix the data layer before evaluating another vendor.
If you'd like help connecting your skills data to your AI strategy, or want me to run a working session with your HR and engineering leaders on hybrid workforce design, I'd love to help.
Sources
- Deloitte, Human Capital Trends (2023-2026), SBO popularization
- Mikaël Wornoo, TechWolf, Future Ready Leadership podcast
- Workday, SAP, Oracle native SBO functionality
- Claude / OpenAI / Cursor / Gemini Skills ecosystem (covered in Writing Skills, Think Big Newsletter #26)