The OECD.AI Policy Navigator

Our policy navigator is a living repository from more than 80 jurisdictions and organisations. Use the filters to browse initiatives and find what you are looking for.

AI Pilot: CV screening in consultants’ recruitment


-
Added by:   OECD analyst
Added on:   23 Sep 2026
Updated by:   OECD analyst
Updated on:   23 Sep 2026

Formez PA's AI Pilot for CV Screening is an internal tool that helps Formez’ selection committees review CVs from candidates applying to become consultants for Formez PA's project activities. It uses natural language processing to identify skills and create a preliminary ranking based on job requirements, reducing manual workload for staff. Launched in May 2025, it addresses the sharp rise in internal applications via the inPA portal while keeping human judgement central to decisions.

Name in original language

Strumento AI per lo screening dei curricula

Initiative overview

The AI Pilot for CV Screening was developed for Formez PA's internal recruitment of consultants who support project activities commissioned to the agency by public administrations.

The introduction of the inPA portal led to a surge in applications: from 11,888 between December 2021 and October 2023 (pre-inPA) to 24,053 for 824 notices from October 2023 to October 2025, a more than 100% increase. This overwhelmed manual screening by three-person committees, requiring 1,500 hours in 2023, 3,000 in 2024, and 2,500 in 2025, with risks of inconsistency and missed talent due to varying committee compositions.

Objectives include automating competency extraction from CVs and preliminary ranking via semantic matching to job needs, while providing dashboards for human review. The tool integrates Formez PA's scoring rules and aims to address aspects related to EU AI Act/GDPR compliance through audits. An interdepartmental team from human resources, project managers, ICT, and end users drives the process.

The initiative will integrate fully into Formez PA's intranet for all internal consultant selections. Future evolution includes refining models via monitoring, potential expansion within Formez PA projects, and capacity-building for staff. Institutionalisation relies on governance, training, and scalable infrastructure to handle growing internal recruitment needs securely.

Results, outcomes and impacts

Since May 2025, the pilot has processed thousands of CVs monthly for internal consultant selections. Besides significantly cutting screening time by about 50% from the 2024 baseline of 3,000 hours, it enhances consistency by benchmarking human versus AI rankings, freeing committees for interviews. Measurements include workload logs, deviation analysis, audit trails, and staff feedback.

Outcomes are fairer internal evaluations and qualitative focus. Future expectations include 20–30% faster cycles for Formez consultant hiring, data-driven improvements to notices, and sustained efficiency gains.

Other relevant details

Challenges: Challenges involve high CV volumes from internal consultant applications, diverse formats risking inaccuracies, evaluation inconsistencies across committees, and addressing EU AI Act/GDPR compliance without bias in rankings. Staff trust in AI and integration with competency models can also be hurdles. Responses feature human-in-the-loop validation, NLP tuning, rotation of selection committee members and Internal Audit checks on winners for declaration accuracy. Audit logs track AI-human deviations. Lessons learned: Lessons highlight early governance, with Formez PA as controller, cross-departmental collaboration, and training for trust. Success requires scalable infrastructure, feedback loops, and clear human oversight communication to support ethical internal recruitment.

About the policy initiative


Category:

  • AI policy initiatives, programmes and projects

Initiative type:

  • AI use cases/projects in the public sector

Status:

  • Active

Start Year:

  • 2025

OECD AI Principles:


Other relevant urls: