What is an internal talent marketplace, and how does it actually work?

An internal talent marketplace is a software platform that matches current employees to internal opportunities based on their skills, aspirations, and availability, rather than on their job title [1][2]. Those opportunities reach past open positions to include short-term projects and gigs, mentorships, rotation and stretch assignments, and learning [1][3][6][10]. Deloitte frames the category as a technology-enabled way to connect employees with opportunities inside the organization, extending beyond full-time roles into gig work, mentorship, rotation programs, and stretch and skill-building experiences [1].

The feature that separates a marketplace from an older internal job board is the skills engine underneath it. Rather than asking people to scroll a list of postings, the platform keeps a skills profile for each employee and a skills requirement for each opportunity, then recommends the matches between them [3][4]. Workday runs this on its Skills Cloud, a machine-learning ontology that it says gives skills a universal language no matter where or how they were gained [4][5]. Fuel50 builds the same idea on an expert-curated skills ontology it calls a skills architecture [3]. The name changes by vendor, though the object underneath stays a living graph of skills, roles, and people.

Why the category exists

Most companies already employ someone who could fill a given opening, yet the default move for a manager with a vacancy is to post it outside [5]. A marketplace exists to make internal candidates visible before that happens, and to put spare skills to work on projects that would otherwise wait for a new hire [1]. Workday cites survey data in which more than half of job seekers had not explored internal roles before looking elsewhere, and more than 80 percent of career moves involved changing employers [5].

Matching runs in two halves: the platform first assembles a skills profile for every employee, then scores that profile against the skills each opportunity requires [3][4]. The quality of the first half decides everything, because a recommendation is only as good as the skills data behind it.

Where the skills data comes from

Three inputs feed a skills profile, and most platforms blend all of them. Employees enter some skills themselves in a profile, the system infers others from work history and HR records, and validation steps confirm the rest through endorsements, assessments, or manager review [4][7]. Workday builds what it calls a skills signature for each worker by inferring and verifying skills derived from its HCM data [4]. ServiceNow describes its Skills Intelligence as using AI to collect, validate, and continuously update employee skills data [7]. Eightfold reads live signals from email, work apps, and tasks to keep what it calls a digital twin of each worker current [6]. Each method leaves a gap: self-reported skills are patchy because people forget or oversell, inferred skills stay guesses until confirmed, and validation costs the time busy managers resist spending.

The skills graph

The engine that makes matching work is a graph that relates skills to one another, not a keyword search. Workday's ontology cleanses skills data by breaking a skill into its components, connecting those components to related skills, and grouping them into categories, so a search for one skill surfaces people who list an adjacent one [4]. ServiceNow applies machine learning to reduce duplicate skills as they are added, since an ungoverned list fills with synonyms that split one capability across three labels [8]. The graph is what lets the system count someone who knows a neighboring tool as a plausible match, not a miss.

Posting and matching opportunities

An opportunity enters the marketplace when a manager or project owner posts it with the skills it needs, and the engine then works in both directions, recommending candidates to the poster and opportunities to employees [2][5]. Workday positions the skills-based recommendations as a way to open conversations between a manager and an employee about internal moves [5]. Eightfold illustrates the output with a worker it says saw an 82 percent match to a role he had never considered, which is the vendor's own example rather than an independent measurement [6].

Full-time role matching and project matching share one skills engine but solve two different problems, and treating them as the same thing is a common setup mistake. Full-time matching is internal recruiting: one person moves permanently into an open role, a manager approves the release, and the marketplace competes directly with the choice to post the job externally [1][5]. The stakes are high and the frequency is low, so this is where manager resistance bites hardest.

Project and gig matching works on a different rhythm. Here an employee keeps their existing job and carves out part of their time for short-term work on another team, which Deloitte describes as letting workers take on projects and tasks anywhere in the organization based on their skills and interests [1][11]. Gloat's Mosaic, for example, breaks a priority initiative into projects and tasks and matches each to the best-fit person rather than filling a permanent seat [2]. Because a gig is temporary and part-time, the risk to a manager releasing someone is smaller, the matches happen far more often, and the skills data improves as people complete work the system can see. Many rollouts start with gigs for that reason, then extend to full-time moves once managers trust the platform. The distinction matters when you compare products, because a marketplace strong at posting permanent roles is not automatically good at slicing initiatives into fractional gigs.

Four issues decide whether a marketplace gets used or quietly dies, and only one of them is about software.

Manager incentives and talent hoarding

The central obstacle is talent hoarding: managers who prevent their best people from moving through blocking, weak references, or simply staying quiet about open roles [12]. The behavior is rational under the usual incentives, because the manager bears close to the full cost of losing a strong performer while the wider organization captures most of the benefit of the move [12]. LinkedIn research cited by 365Talents puts manager resistance as the number one obstacle to internal recruiting for 70 percent of talent professionals, and Deloitte reports that 46 percent of managers resist internal mobility outright [1][12]. Real-time visibility into who is available can reduce the hoarding by making the alternative talent pool obvious, but the durable fix is changing what managers are measured on, not the tool [1].

Skills data quality

A marketplace runs on skills data, so a thin or stale profile produces matches nobody trusts. Self-reported profiles go out of date, inferred skills carry errors until validated, and an ungoverned list fills with duplicates that fragment one capability across several names, which is why ServiceNow builds machine learning specifically to dedupe skills as they arrive [4][7][8]. The profile has to be maintained, and that maintenance competes for the same attention the review cycle and everyday work already claim.

Competing with external hiring and executive sponsorship

Every internal match is a job that did not go to an external requisition, so the marketplace competes with recruiting for the same opening and loses without senior backing. Deloitte names the chief human resources officer as the necessary sponsor and calls the partnership with talent acquisition essential, and it reports that 49 percent of organizations still lack the tools to identify and move people into internal roles [1]. Without that sponsorship, managers post outside by habit and the marketplace stays empty.

Integration with the HRIS and learning systems

The skills profile is only current if it reads from the systems that already hold employee data, and the opportunities are only complete if learning connects to them. Workday keeps its marketplace and Skills Cloud inside the same HCM so the profile draws on the record directly [4], while a standalone tool such as Gloat connects an organization's LXP and LMS so a matched skill gap points at the course that closes it [2]. Cornerstone frames its version as unifying skills across recruiting, learning, and performance for the same reason [9]. A marketplace wired to nothing decays within months.

The skills-based organization puts skills, more than jobs, at the center of how work is organized and how people are deployed, and the internal talent marketplace is the operational engine most vendors attach to that idea [11]. Deloitte describes the model as freeing work from the confines of the job and treating each worker as a portfolio of skills that can move to where the work is, with the marketplace as the mechanism that moves them [11]. The framing is genuine, and it is also where the marketing runs ahead of the evidence. Deloitte's own research finds that only 18 percent of organizations have adopted skills-based approaches to a significant extent, even though 85 percent of HR executives say they plan to redesign work around skills [11]. The vision is real and early, not finished.

Treat vendor outcome numbers with the same caution. Workday reports a 26 percent higher internal hire rate and a 5 percent lower voluntary turnover rate among Skills Cloud customers, and states that 30 percent of the Fortune 500 use it [4]. Those are Workday's figures about Workday's customers, not independent findings, and every vendor in this category publishes similar numbers. They read as a direction of travel rather than proof.

An internal talent marketplace overlaps three neighboring systems, and the clean boundary is which system owns the workflow.

It is not an applicant tracking system

An applicant tracking system runs external hiring: job posts, candidate pipelines, interviews, and offers for people who do not work for you yet. A marketplace runs the internal version of that flow and picks up people already on payroll. The two meet only when an internal candidate applies for a permanent role and moves through a hiring step while staying an employee [1].

It is not a learning management system

A learning management system delivers and tracks training. A marketplace often recommends learning and connects to the LMS, but its job is matching people to work, not hosting courses [2][9]. The learning is the path to an opportunity, not the product itself.

It is not an HRIS

An HRIS is the system of record that holds names, jobs, pay, and reporting lines. A marketplace reads that record to build skills profiles and match people, and it stores no authoritative employment data of its own [4]. When the two share a vendor, as with Workday, the marketplace draws on the record directly; when they do not, it syncs a copy.