Skills Based Workforce Planning and Internal Talent Marketplaces

Skills-based workforce planning is a methodology that reorients staffing models away from fixed headcount-by-role allocations toward a dynamic inventory of discrete, measurable capabilities. Rather than asking "how many agents do we need in Queue X," skills-based planning asks "how many capability-hours of Skill Y must the workforce collectively supply?" The approach has gained significant traction as artificial intelligence systems absorb routine transactional work, leaving human workers to be deployed fluidly across a wider range of contexts. Internal talent marketplaces are the operational mechanism through which organizations enact skills-based planning — matching verified worker capabilities to demand signals in near real time, replacing the assumption that workers are permanently assigned to a single role or queue.

Background and the Limits of Traditional WFM Paradigms

Classical workforce management rests on a headcount-per-queue model. Each queue or workgroup has a forecast, a staffing requirement, a schedule, and an assigned population of agents. The model is intuitive and compatible with the queue-management theory underlying Skill-Based Routing and Erlang-C calculations. It works well when work types are stable, when agents perform one primary function, and when the volume mix between queues is predictable.

These conditions are eroding. Cappelli and Keller (2014) observed that internal labor markets in large organizations had already begun shifting from job-based to project-based assignment structures, driven by the need for rapid capability redeployment across business units. Jesuthasan and Boudreau (2022) argue that AI-driven automation does not eliminate jobs but deconstructs them — disaggregating tasks and reassigning the routine subset to automated systems, leaving humans to manage exceptions, complexity, and judgment-intensive interactions. The practical consequence for workforce management is that the agent-as-queue-member construct becomes an inadequate unit of planning when a single worker may handle escalations from five different automated workflows in a given shift.

Deloitte's 2023 Global Human Capital Trends report identified skills as the new currency of work, noting that 93 percent of surveyed executives considered moving to skills-based approaches important or very important — yet fewer than one in five had done so at scale. The gap between stated priority and operational reality reflects the infrastructure challenge: skills-based planning requires a skills taxonomy, a maintained inventory, a routing or matching mechanism, and a planning model that can consume capability-hours rather than headcount. Most organizations have none of these in place.

Skills Taxonomies and Ontologies

A skills taxonomy is a structured classification of capabilities: organized hierarchically, with defined relationships between skills, and ideally mapped to proficiency levels. Two publicly available frameworks are relevant for WFM practitioners building skills inventories.

ESCO (European Skills, Competences, Qualifications and Occupations) is the European Commission's multilingual classification system covering approximately 13,890 skills and 3,008 occupations, with explicit skill-to-occupation mappings. ESCO distinguishes transversal skills (communication, problem-solving), knowledge concepts, and occupation-specific competencies, making it suitable as a starting point for contact center and knowledge-worker environments.

O*NET (Occupational Information Network), maintained by the US Department of Labor, provides a skills and abilities framework tied to 1,000-plus occupations, including detailed breakdowns of knowledge, skills, abilities, and work activities for roles such as Customer Service Representatives (43-4051) and First-Line Supervisors of Customer Service Workers (43-1011).

In practice, most organizations building WFM-oriented skills inventories develop proprietary taxonomies that overlay a public framework. A contact center taxonomy might organize skills into four domains: (1) channel capability (voice, chat, email, social, video); (2) product or line-of-business knowledge; (3) interaction handling competency (complaint resolution, upselling, technical support); and (4) system proficiency (specific CRM, ticketing, or knowledge base platforms). Each skill is assigned a proficiency scale — commonly three to five levels — with behavioral anchors that define the difference between a Level 2 and Level 3 practitioner.

Proficiency Modeling: Levels, Decay, and Development Velocity

A skills inventory is only operationally useful if it reflects current, accurate proficiency. Two dynamics complicate this: skill decay and development velocity.

Skill decay refers to the erosion of proficiency when a capability is unused. Research on human expertise indicates that skills with procedural components (e.g., operating a specific system, applying a specific protocol) decay faster than conceptual skills. For WFM planning purposes, this means that an agent certified in a skill twelve months ago but not exercised in it since may carry inflated proficiency in the inventory. Skills-based planning systems need either automatic time-decay adjustments or structured re-certification triggers.

Development velocity describes how quickly a worker can acquire a new skill to a defined proficiency level. This varies by skill complexity, worker baseline, training method, and volume of practice opportunities. WFM planners building skills-based capacity models must incorporate realistic ramp curves into projections — a common failure mode is treating a trained agent as immediately equivalent to an experienced practitioner. This connects directly to Handle Time Reduction and AHT Optimization work, where new-to-skill agents consistently produce longer handle times until practice effects accumulate.

Internal Talent Marketplaces

An internal talent marketplace (ITM) is a technology-mediated platform that makes worker capabilities visible across an organization and facilitates dynamic assignment of work — or development opportunities — based on capability match rather than org-chart proximity. The World Economic Forum's 2023 Future of Jobs Report identified internal mobility platforms as among the highest-priority investments for organizations navigating labor market volatility.

ITM mechanics relevant to WFM include:

• Availability signaling — workers indicate capacity or willingness to accept assignments outside their primary queue, enabling planners to identify pools of discretionary capacity.
• Matching algorithms — the platform matches open demand (a project, a queue surge, a coverage gap) to workers whose skills and availability satisfy the requirement. Matching may be rule-based (tiered preference lists) or ML-driven (scoring workers on a composite capability-fit metric).
• Project-based versus permanent assignment — a critical architectural distinction. Traditional WFM assigns workers to queues semi-permanently. ITMs enable time-bounded project assignments — a worker might spend four weeks supporting a product launch queue, then return to their home queue. This changes how Capacity Planning Methods models must account for available capacity at any given time.

In contact center contexts, ITMs reduce the coordination cost of Cross-Training and Skill Mix Strategy by creating a discoverable inventory of cross-trained capability. Without an ITM, cross-trained agents are invisible to planners who do not manage them directly.

Dynamic Skill Routing in a Human-AI Unified Workforce

Traditional Skill-Based Routing assigns contacts to agents based on queue membership and priority rules configured in the ACD. Skills-based routing within an ITM-enabled, human-AI unified workforce operates on a different premise: the workforce includes both human workers and AI agents (chatbots, automated workflows, co-pilot systems), and routing logic must consider the full capability landscape.

In this model, a contact is not assigned to a queue but to the best available capability — which may be a fully automated response, a human with the relevant proficiency, or a human-AI collaborative interaction where AI handles one component while the human handles another. Agentic AI Workforce Planning explores the staffing model implications of this blended environment in detail.

The planning implication is significant: routing decisions and staffing calculations can no longer be modeled as separate queue-specific problems. A worker with skills A, B, and C is simultaneously available to demand streams requiring any of those capabilities, and the optimization problem requires allocating that worker's time across multiple potential demand streams at once. This is fundamentally a capability-hours allocation problem rather than a headcount-per-queue problem.

The Planning Math: From FTE-per-Queue to Capability-Hours-per-Skill

The staffing calculation in a traditional WFM model flows from forecast volume → required service level → Erlang-C calculation → FTE requirement → schedule generation for a specific queue. The unit of output is agents assigned to a queue.

In a skills-based model, the equivalent flow is: forecast demand-per-skill → required capability-hours-per-skill → worker allocation across skill demand streams → schedule generation with multi-skill assignments. The Erlang framework remains applicable at the queue level, but the optimization layer above it must solve a workforce allocation problem across skills simultaneously.

This connects directly to Multi-Skill Scheduling, which addresses the operational scheduling mechanics, and to Workforce Cost Modeling, which must account for the pay and skill-premium structures that typically accompany skills-based labor models. Premium pay for rare or high-value skills is a cost that traditional per-queue FTE models do not capture.

Workforce segmentation becomes an input to this calculation. Workforce Segmentation and Persona Based Planning identifies how different worker populations have different skill profiles and availability patterns — information that feeds the capability-hours supply estimate.

WFM Platform Integration

Major WFM platforms — NICE WFM, Verint Workforce Management, Calabrio ONE — have historically modeled skills as a property of queue assignment rather than a free-standing worker attribute. In these systems, a "multi-skill" agent is an agent assigned to multiple queues, with priority rules governing which queue claims that agent's capacity at a given moment.

Skills-based planning as described here requires a different architecture: skills as a first-class object in the data model, with proficiency levels, decay timestamps, and availability signals. Some platforms have begun adding skills inventory modules or integrating with third-party ITM platforms via API. In environments where the native WFM platform does not support this architecture, organizations typically maintain a skills inventory in an HRIS (Workday, SAP SuccessFactors) or a dedicated skills platform (Gloat, Fuel50, Eightfold), and build integration pipelines to export capability data to the WFM scheduling engine.

WFM Data Infrastructure and Integration Architecture covers the technical requirements for these integration patterns. WFM Technology Selection and Vendor Evaluation provides evaluation criteria for assessing WFM vendor readiness for skills-based scheduling.

Maturity Model Considerations

In the WFM Labs Maturity Model, skills-based workforce planning is primarily an L4–L5 capability. Most organizations at L1–L2 lack the foundational skills taxonomy and proficiency tracking infrastructure required. L3 organizations typically have multi-skill queue assignment in their WFM platform but have not yet modeled capability-hours or integrated an ITM.

Maturity Level	Characteristic
L1–L2	Headcount-per-queue model. Skills are implicit in queue assignment. No formal taxonomy.
L3	Multi-skill queue assignment tracked in WFM platform. Cross-training documented but not systematically inventoried.
L4	Formal skills taxonomy with proficiency levels. Capability-hours planning model. ITM or skills platform in place. Skills-based routing logic in WFM.
L5	Dynamic skill routing across human and AI workforce. Real-time availability signaling. Automated decay tracking and re-certification triggering. Integrated into Agentic AI Workforce Planning architecture.

Related Concepts

• Multi-Skill Scheduling
• Skill-Based Routing
• Cross-Training and Skill Mix Strategy
• Agentic AI Workforce Planning
• Workforce Segmentation and Persona Based Planning
• Workforce Cost Modeling
• Capacity Planning Methods
• WFM Technology Selection and Vendor Evaluation
• WFM Data Infrastructure and Integration Architecture
• Handle Time Reduction and AHT Optimization
• WFM Labs Maturity Model