Skills-Career-Profile Agent
An auditable skills-classification chain from skill extraction to validated career path - every skill mapped to a public taxonomy, every ranking checked for bias, and the final call kept with a person.
Skills-based talent classification: Title VII/Equal Pay Act, EU AI Act Annex III(4)(b) high-risk task assignment, ESCO European Skills + O*NET + Lightcast - GDPR Art. 22 human-in-the-loop.
Analyse your processA selection from over 5,000 projects in 25 years of software development
Can you show that a skills ranking does not quietly disadvantage older employees - and that no opaque model assigned the skill?
The agent maps skills against public taxonomies and runs the classification as mostly deterministic rules - taxonomy mapping to ESCO, O*NET and Lightcast, competency-cluster consistency, the approval workflow - with the AI layer limited to suggesting skill tags, career paths and internal-mobility matches and to a statistical bias check across protected classes. Skill-tag assignment and the final classification stay with people, because GDPR Article 22 reserves the decision for a human and the EU AI Act Annex III(4)(b) classes task-assignment AI as high-risk. The line manager validates, the HR Lead approves, and the model only suggests.
Outcome: The EU AI Act requires a documented rationale for every skills-based assignment recommendation from 2 December 2027, and an opaque ranking creates age-discrimination exposure. US enforcement runs through the EEOC under Title VII, and UK Section 78 Gender Pay Gap reporting applies annually from 250 employees. The agent answers all of this with one auditable chain from skill extraction to a validated career-path simulation.
The architecture reflects that skills-based classification cannot be fully automated but can be consistency-checked across thousands of competency assignments:
Six months of competency-questionnaire ping-pong before the first internal-mobility match - and an opaque skills ranking that quietly disadvantages older employees the whole way.
Skills-matching as compliance trap between Title VII and EU AI Act
This agent follows the Decision Layer principle: each decision is either rule-based, AI-assisted, or explicitly assigned to a human. It is classified per EU AI Act 2024/1689 Annex III(4)(b) as high-risk system - and AI systems for task-assignment based on individual behavior or personal traits are explicitly named in the regulation among the high-risk HR use cases. From 2 December 2027 the agent is therefore subject to enhanced obligations on Risk Management System, Data Governance, Transparency, Human Oversight, and Bias-Audit on skills-ranking patterns.
A typical skills-classification cycle takes six months from skill extraction to internal-mobility communication. The lag between skill emergence on the job and recognized skills-tag in the HRIS is typically 9-14 months - a 40+ employee who acquires a new skill is systematically late in being matched to internal-mobility opportunities versus a younger peer with the same actual skill.
The problem is not the time it takes. It is what happens between the steps: no systematic equity analysis across protected classes, no documented rationale for a Title VII challenge, no competency-cluster consistency check, no auditable taxonomy reference, and no defensible audit trail when an EEOC or EHRC pattern-and-practice investigation arrives.
EU AI Act Annex III(4)(b) task-assignment-personal-traits high-risk
The Skills-Career-Profile Agent falls under EU AI Act 2024/1689 Annex III Point 4 Letter b - the regulation explicitly lists AI systems for task-assignment based on individual behavior or personal traits as high-risk HR AI use cases. The obligations from 2 December 2027 include:
- Article 9 Risk Management System: identifying, assessing and mitigating bias risks in skills-ranking patterns across gender, age, ethnicity and disability
- Article 10 Data Governance: training-data quality and bias detection through demographic-parity and equal-opportunity tests, anchored in the public ESCO, O*NET, Lightcast and UK NQF taxonomies
- Article 13 Transparency Obligations: documentation of how the system works, its accuracy and robustness, and its bias-audit results
- Article 14 Human Oversight: mandatory human-in-the-loop on every skills-classification recommendation
- Article 26 Deployer Obligations: a DPIA, supervisory-authority consultation, post-market monitoring and incident reporting on skills-bias incidents
- Article 27 FRIA Fundamental Rights Impact Assessment: only for public bodies and private entities providing public services, before deployment, in consultation with the DPO and employee representation bodies (such as works councils) or unions
Fines up to 15M EUR or 3 percent global group revenue. Cross-Reference EU GDPR Article 35 DPIA obligation in systematic automated assessment with significant impact on the employment relationship.
US Title VII, the Equal Pay Act and the UK Equality Act
Title VII (42 USC 2000e) prohibits skills-based assignment discrimination on the basis of race, colour, religion, sex or national origin, and the Equal Pay Act (29 USC 206(d)) applies to the resulting pay change when skills-classification drives compensation banding. The Lilly Ledbetter Fair Pay Act of 2009 restarts the statute of limitations with each new paycheck, so an undervalued classification or a missed skill tag can be litigated years later, once subsequent paychecks reveal the pattern.
ADEA Age Discrimination in Employment Act 1967 (40+ years protected) applies at skills-based internal-mobility screening, including skills-classification ranking when AI software systematically deprioritizes 40+ candidates in emerging skills clusters. ADA Americans with Disabilities Act applies to competency-profile accommodation. GINA Genetic Information Nondiscrimination Act prohibits genetic-information-based skills-tagging. EEOC Algorithmic Discrimination Initiative 2021-2026 explicitly targets AI-driven skills-classification screening as a top enforcement priority.
The UK Equality Act 2010 protected characteristics apply at every skills-classification decision, and the equal-pay provisions consolidated into Sections 64 to 66 apply to the comparator analysis behind the resulting pay change.
US state pay transparency and UK Section 78 Gender Pay Gap
More than 25 US state and city pay-transparency laws now reach skills-driven internal-mobility postings. California SB 1162 requires a pay range on the posting and three years of records for the candidates. Colorado goes furthest, requiring proactive notice of every internal-mobility opportunity to all employees. EEOC enforcement of EEO-1 Component 2 demographic-pay reporting now covers these skills-driven changes as well, and UK Section 78 makes gender pay gap reporting mandatory each year from 250 employees.
UK Section 78 Gender Pay Gap Reporting Regulations 2017 require mandatory annual reporting 250+ UK employees by 4 April - EHRC has been increasingly focused on skills-based internal-mobility gap indicators because skills-classification patterns drive much of the structural pay gap. ACAS Code of Practice provides procedural guidance. UK ICO enforces UK GDPR Article 22 automated decision-making prohibition. Gender pay gap data are published on the government’s service, and missing or non-compliant figures invite EHRC enforcement.
The ESCO, Lightcast and O*NET taxonomies, and the risk of AI bias
The skills-taxonomy backbone rests on four publicly maintained classifications. ESCO, from the European Commission, provides 13,944 skills and 3,008 occupations crosswalked to ISCO-08 across 27 EU languages. O*NET, from the US Department of Labor, provides over 1,000 occupations with their skills, abilities and knowledge. The Lightcast Open Skills Library provides over 32,000 open skills with machine-readable definitions, and the UK NQF supplies eight qualification levels under Ofqual regulation. Stable concept URIs make cross-system referencing auditable.
The risk vectors of AI bias in skills-fit ranking are familiar - training data that underrepresents older employees, proxy variables for age such as LMS recency driving the ranking, feedback loops, and opacity when an employee cannot challenge a ranking under GDPR Article 22(3). The mitigations are a quarterly bias audit with demographic-parity and equal-opportunity tests, a documented rationale for the Title VII audit trail, and a DPIA before deployment.
Cross-reference to Promotion-Process, Performance-Review-Documentation and Succession-Planning
The Skills-Career-Profile Agent sits in a pipeline of specialised HR agents. The Performance-Review-Documentation Agent supplies the ratings and 360 feedback used as skills-extraction input. The Promotion-Process Agent reuses the Skills Equity Analysis Engine for promotion equity analysis. The Compensation-Benchmarking Agent receives the validated competency profiles for compensation banding, and the Payroll-Reporting Agent generates the CSRD and ISO 30414 figures. The HR-Document-Management Agent archives the Title VII rationales, and the Audit-Compliance Agent verifies the EU AI Act deployer obligations on the skills-classification software.
At a glance
- Classification: EU AI Act 2024/1689 Annex III(4)(b) high-risk HR Task-Assignment-Personal-Traits explicitly named in the regulation, applicable from 2 December 2027
- Compliance anchors: Title VII and the Equal Pay Act in the US, the UK Equality Act and Section 78, and the EU AI Act, with GDPR Article 22 and CSRD ESRS S1-12 training and skills development reporting
- Taxonomy reference: ESCO (13,944 skills across 27 EU languages), O*NET (over 1,000 occupations), the Lightcast Open Skills Library (over 32,000 open skills) and the UK NQF (8 levels)
- Codetermination: German Section 87(1) No. 6 works council co-determination for the IT system
- Equity threshold: an unexplained gender or age skills-classification gap above 5 percent triggers a Joint Pay Assessment and a six-month remediation duty
- Fines: up to 15M EUR or 3 percent of global group revenue under the EU AI Act, up to 4 percent under GDPR, on top of Title VII class-action damages and UK EHRC enforcement
- Audit obligation: a DPIA and a quarterly bias audit, with auditor verification under the CSRD and the UK Section 78 annual report
Decision-Maker Distribution Skills-Career-Profile
| Step | Decider | Rationale |
|---|---|---|
| Skill extraction | A | NLP extraction from reviews, projects and LMS records |
| Taxonomy mapping | R | ESCO, O*NET, Lightcast and UK NQF |
| Competency-cluster validation | R | Role architecture and Hay Points |
| Self-assessment intake | H | Individual employee claim with evidence |
| Manager validation | H | Line manager evidence review and proficiency confirmation |
| Skills equity analysis | A | ML-statistical bias detection across protected classes |
| Career-path simulation | A | Model-suggested next role and skills gap |
| Internal-mobility match | A | ML-supported skills-fit ranking against requisitions |
| Equity escalation | R | Threshold >5 percent EU Pay Transparency 2023/970 |
| Works council co-determination | R | German Section 87(1) No. 6 IT-system co-determination |
| HR Lead approval | H | Final approval Title VII-compliant |
| EU GDPR information | R | Article 13+14+22 (3) standard workflow |
| Skills-profile documentation | A | Model-drafted profile and career-development plan |
| CSRD/ISO reporting | R | ESRS S1-12 training and skills development, ISO 30414 |
Micro-Decision Table
Who decides in this agent?
14 decision steps, split by decider
Extract raw skill signals from reviews, projects and learning records Which raw skill signals are extracted from the performance reviews, project history, LMS learning records, 360 feedback and certification records? AI Agent Employee
The model extracts skill signals from performance reviews, project history, LMS records and certifications, suggesting skill tags. Its output requires human validation under EU AI Act Article 14; the downstream taxonomy mapping follows a deterministic rule.
Decision Record
Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.
Challengeable by: Employee
Map the skills to the ESCO, O*NET and Lightcast taxonomies Are the extracted skill tags mapped to stable concept URIs in ESCO, O*NET and the Lightcast Open Skills Library, with UK NQF level mapping? Rules Engine Auditor
A deterministic rule maps each extracted tag to a stable concept URI in ESCO, O*NET and Lightcast, with UK NQF level mapping. These public taxonomies give the auditable data-governance reference the EU AI Act Article 10 requires.
Decision Record
Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.
Challengeable by: Auditor
Validate the competency clusters against the role architecture Are the mapped skills assigned to competency clusters consistent with the role architecture and the Korn Ferry Hay methodology? Rules Engine
A deterministic rule checks the mapped skills against the competency framework and role architecture, flagging any skill-cluster gap for line manager review.
Decision Record
Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.
Take in the employee's self-assessment with its evidence Which self-assessed skills are accepted, with a proficiency-level claim backed by evidence such as project records, certifications or manager validation? Human
The employee makes the skill claim and supports it with evidence - project records, certifications or manager validation - for the Title VII audit trail. Any AI suggestion is a recommendation, not the decision, and EU AI Act Article 14 makes this human oversight mandatory.
Decision Record
Challengeable: Yes - via manager, works council, or formal objection process.
Have the manager validate the skill claim and proficiency level Does the line manager validate the employee's skill claim, confirming the proficiency level after reviewing the evidence? Human
The line manager validates the skill claim against project context, observed performance and the supporting evidence, confirms the proficiency level and documents the rationale for the Title VII audit trail. EU AI Act Article 14 makes this human oversight mandatory.
Decision Record
Challengeable: Yes - via manager, works council, or formal objection process.
Run the skills equity analysis for Title VII disparity indicators Are there systematic skills-classification disparities across the protected classes - gender, age, ethnicity, disability and the others - above the 5 percent threshold of the EU Pay Transparency Directive and the Title VII statistical-significance test? AI Agent Auditor
The model runs a statistical analysis for skills-classification disparities across protected classes under Title VII and the UK Equality Act. Its output is an indicator, not a final decision, and is validated by the HR Lead, DPO and EEOC compliance officer.
Decision Record
Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.
Challengeable by: Auditor
Simulate the career path, next role and skills gap Which next-role suggestions and required-skills-gap simulations are generated for the employee's career path? AI Agent Employee
The model simulates next-role suggestions and the skills gap to reach them, with a recommended learning path. Its output is an indicator, not a final decision, validated by the employee, line manager and HR Business Partner, with a documented rationale per suggestion.
Decision Record
Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.
Challengeable by: Employee
Match the skills profile to open requisitions for internal mobility How is the employee's skills profile matched against the skills requirements of open requisitions for internal-mobility ranking? AI Agent Employee
The model ranks the employee's skills profile against open requisitions. Its output is an indicator, not a final decision, validated by the hiring manager and HR Business Partner with a documented rationale. An opaque skills-fit ranking is itself a Title VII risk, which the design guards against.
Decision Record
Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.
Challengeable by: Employee
Escalate a skills-equity finding under the EU AI Act deployer duties Is a skills-equity issue escalated to the HR Lead, executive leadership and works council under the EU AI Act Article 26 deployer obligations? Rules Engine
A deterministic threshold escalates any unexplained gender or age skills-classification gap above 5 percent, feeding the EU Pay Transparency remediation duty, US EEOC enforcement and UK Section 78 reporting.
Decision Record
Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.
Notify employee representation bodies (such as works councils) and run the co-determination procedure on the skills software Is the employee representation involved where applicable - the works council's co-determination under the German Works Constitution Act Section 87, or US NLRA collective bargaining? Rules Engine
A deterministic rule runs the works council procedure where co-determination applies and records the rationale. In Germany, the skills software needs the works council's consent under Section 87(1) No. 6; if they cannot agree, the conciliation board decides.
Decision Record
Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.
Confirm the HR Lead's final approval of the skills profile Is the validated skills-career profile given final approval by the authorised HR leader, with a Title VII-compliant audit trail? Human
An authorised HR leader gives the final approval with a documented Title VII rationale. GDPR Article 22 bars a solely automated decision and EU AI Act Article 14 makes this human oversight mandatory.
Decision Record
Challengeable: Yes - via manager, works council, or formal objection process.
Notify employees of the classification and right to challenge How are employees transparently informed of their skills classification and career-path simulation, and of the right to challenge it, under the GDPR Article 13 and 14 duties, Article 22(3) and EU AI Act Article 13? Rules Engine Employee
A deterministic workflow generates the standardised notification with a rationale block, meeting the GDPR Article 13 and 14 information duties and the Article 22(3) right to challenge an automated recommendation. Opaque communication about the classification rationale is itself a Title VII risk.
Decision Record
Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.
Challengeable by: Employee
Generate the skills-profile documentation and hand over the learning path Are the skills profile, career-development plan, recommended learning paths and competency-gap roadmap generated and handed over to the Learning-Path-Recommendation and Performance-Review-Documentation agents? AI Agent
The model generates the skills-profile summary, career-path simulation and competency-gap roadmap from the approved data. Its output requires human validation under EU AI Act Article 14; distribution follows a deterministic rule.
Decision Record
Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.
Update the CSRD and ISO 30414 reporting with the skills figures How is the skills classification documented for the CSRD skills-development, ISO 30414, EU Pay Transparency and UK Section 78 reporting? Rules Engine
A deterministic rule generates the skills-development figures for CSRD ESRS S1-12 and the UK Section 78 report due 4 April, drawing on the ISO 30414 skills and capabilities metrics.
Decision Record
Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.
Decision Record and Right to Challenge
Every decision this agent makes or prepares is documented in a complete decision record. Affected employees can review, understand, and challenge every individual decision.
Does this agent fit your process?
We analyse your specific HR process and show how this agent fits into your system landscape. 30 minutes, no preparation needed.
Analyse your processGovernance Notes
Legal basis: ESRS numbers follow Delegated Regulation (EU) 2026/1563, which applies from FY2027 and may be chosen for FY2026; in the 2023/2772 version, the disclosures from S1-3 onwards carry the next-higher number. Under Directive (EU) 2026/470, the CSRD covers companies with more than 1,000 employees and more than EUR 450 million net turnover from FY2027; the first wave, large public-interest entities with more than 500 employees, reports for FY2024 to FY2026. Legal status: September 2026.
Assessment
Prerequisites
- ESCO + O*NET + Lightcast Open Skills Library + UK NQF taxonomy reference + concept URIs
- Role architecture + competency-cluster framework + Hay Points methodology
- Performance ratings + 360 feedback from Performance-Review-Documentation-Agent
- LMS records + certification database + project tracking system integration
- Approval matrix by skills-cluster + line manager + HR Business Partner + HR Lead
- Works council/Union agreement on AI-supported skills-classification governance
- DPIA documentation EU GDPR Article 35
- EU Pay Transparency-compliant comparator group definition >=6 employees per skills-cluster
- Bias-Audit framework with quarterly Demographic Parity tests
Infrastructure Contribution
What this assessment contains: 9 slides for your leadership team
Personalised with your numbers. Generated in 2 minutes directly in your browser. No upload, no login.
- 1
Title slide - Process name, decision points, automation potential
- 2
Executive summary - FTE freed, cost per transaction before/after, break-even date, cost of waiting
- 3
Current state - Transaction volume, error costs, growth scenario with FTE comparison
- 4
Solution architecture - Human - rules engine - AI agent with specific decision points
- 5
Governance - EU AI Act, works council, audit trail - with traffic light status
- 6
Risk analysis - 5 risks with likelihood, impact and mitigation
- 7
Roadmap - 3-phase plan with concrete calendar dates and Go/No-Go
- 8
Business case - 3-scenario comparison (do nothing/hire/automate) plus 3×3 sensitivity matrix
- 9
Discussion proposal - Concrete next steps with timeline and responsibilities
Includes: 3-scenario comparison
Do nothing vs. new hire vs. automation - with your salary level, your error rate and your growth plan. The one slide your CFO wants to see first.
Show calculation methodology
Hourly rate: Annual salary (your input) × 1.3 employer burden ÷ 1,720 annual work hours
Savings: Transactions × 12 × automation rate × minutes/transaction × hourly rate × economic factor
Quality ROI: Error reduction × transactions × 12 × EUR 260/error (APQC Open Standards Benchmarking)
FTE: Saved hours ÷ 1,720 annual work hours
Break-Even: Benchmark investment ÷ monthly combined savings (efficiency + quality)
New hire: Annual salary × 1.3 + EUR 12,000 recruiting per FTE
All data stays in your browser. Nothing is transmitted to any server.
Skills-Career-Profile Agent
Initial assessment for your leadership team
A thorough initial assessment in 2 minutes - with your numbers, your risk profile and industry benchmarks. No vendor logo, no sales pitch.
All data stays in your browser. Nothing is transmitted.
Related Pages
Agent Blueprint Available
A full blueprint for Skills-Career-Profile Agent is available with micro-decision decomposition, industry variants, and implementation details.
View BlueprintRelated Agents
People Analytics Agent
Workforce intelligence - from attrition prediction to engagement drivers.
Performance Review Documentation Agent
Structured review documentation - consistent, complete, and audit-ready.
Promotion-Process Agent
An auditable promotion chain from nomination to title change - every band jump validated, every pay-equity gap surfaced, and the nomination and final approval kept with people, not the model.
Frequently Asked Questions
Does the agent make autonomous skills-classification decisions?
Why is this agent an EU AI Act high-risk system under Annex III(4)(b)?
How does mapping to the ESCO, O*NET, Lightcast and UK NQF taxonomies make the classification auditable?
How is skills equity ensured under Title VII, the UK Equality Act and the EU Pay Transparency Directive?
How is the risk of AI bias mitigated for skills-classification software?
What cross-references to other HR agents exist?
What Happens Next?
30 minutes
Initial call
We analyse your process and identify the optimal starting point.
1 week
Discover
Mapping your decision logic. Rule sets documented, Decision Layer designed.
3-4 weeks
Build
Production agent in your infrastructure. Governance, audit trail, cert-ready from day 1.
12-18 months
Self-sufficient
Full access to source code, prompts and rule versions. No vendor lock-in.
Implement This Agent?
We assess your process landscape and show how this agent fits into your infrastructure.
Analyse your HR process