How It Works
The Internal Mobility Matchmaker is a two-sided AI tool that brings together hiring managers and employees on a single platform. Instead of posting roles externally before checking internal talent, managers can instantly surface the best internal candidates. And instead of waiting for job postings to appear, employees can proactively discover where their skills are the strongest fit across the organization.
Manager Mode
A manager defines an open role by entering the title, department, salary range, and required skills. The AI evaluates all 25 employees in the org and returns the top 5 internal candidates, ranked by match score. Each candidate card shows:
- Match score (0–100) with a color-coded bar
- Fit label: Great Fit (75+), Good Fit (50–74), or Poor Fit
- Pay increase estimate based on the salary range midpoint
- Skills matched (green chips) and skills gap (amber chips)
- AI reasoning — a short narrative explaining why this person fits the role
Employee Mode
An employee selects their profile from a dropdown and sees a preview card showing their current role, salary, performance rating, tenure, and skills. They can optionally add additional skills not yet on file. The AI evaluates all 15 open positions and returns the top 5 best-fit matches, each showing:
- Match score with a color-coded fit label
- Salary change (positive or negative) relative to current compensation
- Skills you already have (green) and skills to grow (amber)
- AI reasoning explaining why this role fits their background
The Data
The demo uses a realistic fictional org of 25 employees across Engineering, Sales, Marketing, Operations, Finance, HR, Customer Success, and Product, matched against 15 open job postings ranging from mid-level to director. All names, salaries, and data are fabricated for demonstration purposes.
Bias-Aware AI Matching
The tool ships with two complementary, default-off bias-control toggles. Together they cover both layers of LLM bias risk in AI-assisted hiring decisions:
- Anonymize sensitive context for AI matching — strips names and school affiliations from the candidate data before it reaches the AI. AI models can inadvertently use these as proxies for protected demographic attributes (race, gender, national origin, age). When enabled, those fields are decoded back from the candidate ID for display only.
- Apply EEOC-aligned anti-bias rules — prepends an explicit instruction set inside the prompt: ignore protected-class attributes (race, gender, age, disability, religion, national origin, marital/parental status, etc.) and their proxies, avoid "cultural fit" reasoning, and don't penalize career gaps without role-relevant evidence. Aligns with U.S. EEOC, UK Equality Act, EU Equal Treatment, and Canadian Human Rights frameworks.
Both toggles can be enabled together (belt + suspenders) or independently, depending on the user's compliance posture and jurisdiction.
Business Value
Internal mobility programs reduce recruiting costs, improve retention, and surface hidden talent. Most organizations struggle to match internal candidates to openings systematically — this tool demonstrates how AI can automate that process at scale, giving HR and people analytics teams a practical, cost-effective starting point.
From Demo to Production
This demo uses a fictional 25-person org. A production deployment connects to your HRIS and ATS to match your real workforce against your real open roles — continuously.
Real-World Challenges
| Challenge | Why It's Hard |
|---|---|
| Data completeness | Employee skill profiles are often outdated or incomplete — people gain skills faster than HR systems track them. |
| Manager resistance | "I want to hire external talent, not lose my best people to another team." Internal mobility requires cultural buy-in, not just tooling. |
| Confidentiality | Employees browsing internal roles don't want their manager to know. The system must support discreet exploration. |
| Skill taxonomy | Standardizing skills across departments that use different terminology (e.g., "client management" vs "account management") requires a shared ontology. |
| Career path sensitivity | Suggesting lateral moves vs promotions requires nuance. A "match" that feels like a demotion damages trust in the tool. |
| Headcount planning integration | Matching must respect approved headcount and budget, not just surface any open req. |
Cost Estimates
| Component | Starter | Growth | Enterprise |
|---|---|---|---|
| AI API (GPT-4o-mini / GPT-4o) | $30–100/mo | $100–400/mo | $400–1,500/mo |
| HRIS integration (Workday, BambooHR) | $200–600/mo | $600–2,000/mo | $2,000–8,000/mo |
| ATS integration (Greenhouse, Lever) | $100–400/mo | $400–1,200/mo | $1,200–4,000/mo |
| Employee self-service portal | $0–200/mo | $200–800/mo | $800–3,000/mo |
| Total monthly | ~$300–1,000 | ~$1,000–4,000 | ~$4,000–15,000 |
ROI Definition
- Primary metric: Reduction in external hiring for backfillable roles (target: 15–30% of roles filled internally that would have gone external)
- Secondary metric: Retention improvement — employees who see internal mobility options stay 2x longer
- Break-even: Typically within 2–3 months
- Concrete example: Average external hire costs $15K (recruiting + onboarding). Filling 20 roles internally per year instead of externally = $300K saved vs ~$40K/year platform cost
Build It or Buy It?
A note on pricing, up front. No priced comparison appears here, deliberately. The talent-marketplace category — Gloat, Eightfold, Fuel50 — is entirely quote-only, and pricing in this space varies by headcount, module and contract term in ways no published figure would capture. We would rather omit the numbers than print ones we cannot cite. The more useful observation is that the constraint on this project is almost never budget — it is whether you have usable skills data, which the sections below cover.
What building it actually requires — skills, systems and the ongoing work
The hard problem is not matching. It is knowing what your people can do.
Matching a skills profile to a role description is the easy part and largely what this demo shows. Every organisation that attempts internal mobility discovers the same thing: the skills data does not exist, and what does exist is unreliable.
| Source | Why it disappoints |
|---|---|
| Job titles in the HRIS | Describe a position, not a capability, and are often years stale. Two people with identical titles may share almost no skills. |
| Self-reported skills | Aspirational and unevenly reported. Confidence varies by personality and, measurably, by demographic group — so a self-report system quietly advantages whoever is most comfortable self-promoting. |
| Manager assessments | Rate performance in the current role, which is not the same as capability for a different one, and carry the manager's own incentive to retain good people. |
| Inferred from work artefacts | The most accurate and the most invasive. Deriving skills from tickets, commits, documents or CRM activity raises real employee-monitoring and consent questions that need answering before, not after. |
Skills you need on hand
| Area | Why it is needed |
|---|---|
| A skills taxonomy | Somebody must decide that "data analysis", "analytics" and "reporting" are or are not the same capability. Without a controlled vocabulary, matching degrades into string similarity. |
| HRIS and ATS integration | Reading the person, reading the open role, and writing back an expression of interest so it becomes an actual application rather than a suggestion nobody actions. |
| Adverse-impact monitoring | An internal mobility engine decides who is shown opportunities. That is a selection process, and it needs the same fairness scrutiny as external hiring. |
| Explainability | An employee shown a recommendation will ask why — and so, eventually, will someone reviewing whether opportunities were distributed fairly. |
The part that is easy to underestimate: manager hoarding is the actual blocker, and it is not technical. Managers are measured on their own team's delivery and lose when a strong performer moves internally. Unless that incentive changes, the tool will surface matches that quietly never progress, and the failure will be invisible — it looks like low engagement rather than a structural conflict. Every internal mobility programme that works has addressed this, usually before it bought any software.
The second thing: recommending opportunities is a decision with legal exposure. If the system systematically under-surfaces roles to a protected group, you have built something that limits opportunity at scale, with an audit trail. US EEOC principles on selection procedures apply to internal movement as well as external hiring, and the EU AI Act treats AI used for promotion decisions as high risk. Measure selection rates by group from day one.
Choose build when you already hold genuine skills signal — a completed skills-assessment programme, a well-maintained competency framework, or defensible inference from work systems — and you want it wired into your own HRIS and ATS.
Choose a talent marketplace when you need the taxonomy, the employee-facing experience and the change-management playbook as much as the matching. Those vendors sell a programme, not an algorithm, and the programme is the part that determines whether it works.
What to expect if you go ahead — timeline, adoption, and where it goes wrong
Timeline. Matching profiles to roles works almost immediately. Assembling skills data anyone trusts takes months, and it is the project. Any plan that treats data collection as a preliminary step has the proportions backwards.
Adoption is the real risk, not accuracy. These systems fail quietly: employees look once, see recommendations that feel generic or unattainable, and never return. A small number of well-matched, genuinely available opportunities beats comprehensive coverage of roles nobody can realistically move into.
Where it actually goes wrong
- Recommendations that go nowhere. An employee expresses interest and hears nothing. One round of that ends participation for that person permanently, and word travels.
- Stale role data — positions already filled, or filled internally by someone's existing network before ever appearing in the system.
- Title-matching in disguise. Without a taxonomy, the system recommends the same job the person already has at a different business unit.
- The confidence gap gets encoded. Self-reported skills favour people comfortable claiming expertise, so a system built on them amplifies an existing disparity rather than correcting it — the precise opposite of the stated goal.
- Manager veto in practice but not in policy. Movement is nominally supported and consistently blocked, and the system's metrics show engagement without outcomes.
The honest question to ask first: if an employee applies for an internal role today, what actually happens? If the answer is "it depends on their manager" or "there is no process", software will not fix that — it will surface the problem faster and to more people. Fix the pathway first, then build the thing that points people down it.
Sourcing note. No prices are quoted in this section, deliberately. Gloat, Eightfold and Fuel50 do not publish list pricing, and we would rather omit a number than print one we cannot cite. The cost estimates earlier on this page are SkillEra's own build-side modelling, not vendor list prices. The regulatory observations above are general and not legal advice — take advice on your own jurisdictions before deploying automated recommendation of internal opportunities.
Technology Stack
- AI Model: OpenAI GPT-4o-mini
- Backend: Next.js API route (serverless)
- Frontend: React client with dual-mode interface (Manager + Employee)
- Data: Synthetic employee and role dataset (25 employees, 15 open roles)
Want This for Your Business?
Connects to your HRIS and ATS for real-time internal mobility matching. Includes manager dashboard, employee self-service portal, and skills gap reporting. A full deployment typically takes 3–5 weeks and starts at $5,000.