About this research
This study examines how career guidance practitioners integrate generative AI tools — such as ChatGPT, Claude, and Perplexity — into their daily work with adult clients. It is the first systematic empirical study of this topic in the Italian context, and one of the very few internationally.
The research adopts a mixed-methods design based on three instruments: an online survey administered to 81 career guidance practitioners; semi-structured in-depth interviews with 9 practitioners from Northern Italy who already use generative AI regularly in their practice; and a quantitative questionnaire assessing the perceived usefulness of AI tools across standardised counselling tasks, completed by 7 of the 9 interviewees. The fieldwork was carried out in June and July 2025.
Why this study matters
Generative AI platforms are qualitatively different from previous digital tools: they can perform directly some of the most cognitively demanding tasks in guidance practice, including synthesising client profiles, identifying occupational pathways, generating personalised documents, and simulating job interviews. Systematic reviews by Bankins et al. (2024) and Pandya & Wang (2024) have confirmed the near-total absence of empirical research on how practitioners actually use these tools with adult clients.
Key findings
Adoption is still limited. 69% of survey respondents either do not use generative AI at all or use it with fewer than 10% of their clients. The factors most associated with non-use are age over 50, more than ten years of experience, and employment in public employment centres. The main barrier is insufficient skills and practical knowledge; the main enabler would be structured training with hands-on practice.
Eight functions, 74 activities. The interviews generated a taxonomy of eight distinct functions through which AI is used in guidance practice: Author/Editor (drafting CVs, cover letters, competence reports); Coherence Analyst (matching client profiles to occupational pathways); Intelligent Search Engine (researching employers, sectors, training options); Sparring Partner (simulating job interviews); Strategic Co-adviser (developing action plans); Practitioner Supervisor (back-office case review); Professional Learning Tool; and Intervention Designer. The first two functions are used by all nine interviewees and account for 58% of coded activities.
Two modes of integration. The study identifies a co-pilot mode — fast, task-focused, episodic — and a co-thinking mode, in which the practitioner engages in iterative dialogue with the AI, activating a metacognitive process that improves the overall quality of the counselling session. This distinction, not previously described in the literature, has direct implications for training.
The practitioner role is being reconfigured. The dyadic practitioner–client relationship becomes triangular: practitioner – client – AI. Expertise shifts from content production to output validation, from information provision to prompt engineering and critical interpretation, from expert authority to guided sense-making. Eight of nine practitioners already teach clients to use AI tools directly.
Three critical issues. 67% of interviewees use personal rather than organisational accounts when processing client data, in violation of GDPR. Eight of nine organisations have introduced no shared guidelines, no training, and no structural changes in response to the AI transition. And the type of guidance received is increasingly determined by clients’ digital literacy, risking a two-tier service.
Download the Executive Summary and other documents
The Executive Summary (approx. 8 pages) presents the full research design, findings, and implications for policy and practice.
- Download Executive Summary – PDF 2026-5-8-executive_summary
- Download Between Mercury and Virgil: Metaphors for Generative AI in Career Guidance Practice PDF 2026-5-8-metaphors
- Download Teaching Generative AI to Career Guidance Clients: Digital Autonomy, Disintermediation, and the Changing Role of the Practitioner PDF 2026-5-8-disintermediation
- Download Eight Functions of Generative AI in Career Guidance: Evidence from Italian Practitioners PDF 2026-5-8-functions
- Download Generative AI in the Career Guidance Session: Dynamics, Role Reconfiguration, and Professional Competence PDF 2026-5-8-interview
- Download Grassroots Innovation, Organisational Inertia: Generative AI in Italian Career Guidance Organisations PDF 2026-5-8-organization
About the author
Leonardo Evangelista is an independent researcher and career guidance practitioner with over twenty years of experience in the field. He is the founder of orientamento.it (this website), one of Italy’s leading career guidance resources.
ORCID: https://orcid.org/0009-0009-7855-4023 Contact: l.evangelista@orientamento.it