UNI 11621-8:2026: the 12 AI Professional Profiles

On April 30, 2026, the UNI 11621-8:2026 standard was published, developed by UNI with coordination from the Digital Transformation Department of the Presidency of the Council of Ministers. It is the first national standard in Europe to systematically define the professional profiles of Artificial Intelligence: 12 roles, each with measurable skills, responsibilities, and performance indicators (KPIs). An operational reference for companies, public administrations, training bodies, and certification organizations, in line with the AI Act and Law 132/2025.

What the UNI 11621-8:2026 standard is

Until April 30, 2026, in Italy there was no shared national reference for defining what it means to work professionally in the AI sector: which skills, which responsibilities, which expected outcomes. The market has developed rapidly and, as noted by those who worked on the standard, in a context with still-ambiguous definitions, professional roles and titles have multiplied in a largely uneven way.

The UNI 11621-8:2026 standard does not solve everything at once. But it establishes a fixed point: it defines 12 professional profiles, specifying for each one the mission, main tasks, expected outcomes, skills, knowledge, abilities, and key performance indicators (KPIs).

It was developed by UNI Technical Committee UNI/CT 526 – UNINFO, with the coordination of the Digital Transformation Department of the Presidency of the Council of Ministers and with the contribution of UNI Technical Committee 533 “AI”. The methodology adopted is the one already consolidated in the UNI 11621 series, aligned with the European e-Competence Framework (UNI EN 16234-1) and started in 2021 with seven parts dedicated to ICT profiles:

Part 1 (2021): Methodology for building profiles based on e-CF
Part 2 (2021): Second-generation European ICT profiles
Part 3 (2021): Professional profiles operating on the web
Part 4 (2024): Profiles related to information security
Part 5 (2022): Profiles related to geographic information
Part 6 (2021): Profiles related to ICT metrics management
Part 7 (2024): Profiles related to digital transition

Part 8 expands this corpus into the AI sector, using the same methodology and the same logic. Those already familiar with the series will find a coherent approach; those encountering it for the first time will find a standard already embedded in a mature regulatory system.

The 12 professional profiles

The standard identifies 12 roles that cover the entire lifecycle of AI systems — from strategic governance to research, including development, data management, security, and solution design:

Chief AI Officer
AI Consultant
AI Product Manager
AI Prompt Engineer
AI Algorithm Engineer
AI Deep Learning Engineer
AI Data Engineer
AI Data Scientist
AI Security Specialist
AI Machine Learning Engineer
AI Natural Language Processing Engineer
AI Research Scientist

For each one, the standard defines not only the required competencies, but also the expected level of autonomy and responsibility and the indicators used to measure the work performed concretely. One detail worth highlighting: the standard applies to those who design, develop, integrate, and manage AI systems. It does not concern people who simply use existing AI tools.

The standard also gives explicit attention to governance, risk management, security, transparency, explainability, and regulatory compliance, with reference to the AI Act, GDPR, and the international standard UNI CEI ISO/IEC 42001.

The short descriptions of the individual profiles are editorially prepared. The full standard is available on the UNI store.

Why now

The timing is no coincidence. Europe is building the implementation rules of the AI Act, Regulation (EU) 2024/1689, which explicitly requires that artificial intelligence systems be developed and managed by people with adequate skills. The issue is that the Regulation establishes this obligation without defining operationally what it means: it sets a requirement, but does not provide the tools to verify it.

The UNI 11621-8:2026 standard fills this gap with a voluntary technical standard that makes a European requirement concrete. By integrating with Law No. 132 of September 23, 2025, which promotes AI literacy and certification of AI skills, it translates a European regulatory obligation into a set of verifiable roles and competencies, useful to businesses, public administrations, training bodies, and certification organizations.

Undersecretary Alessio Butti commented on the publication with these words: “We are the first in Europe to have released a standard, consistent with the AI Act and national legislation, that aims to strengthen the scope of competencies and related responsibilities in artificial intelligence. The 12 profiles defined by the standard represent an operational tool for businesses, public administrations, and the education system, because they make it possible to qualify and certify competencies in a uniform way.”

Italy is positioning itself as a European first mover. According to UNI and the Italian government, no other European country, including France and Germany, which both have active national AI programs, has yet published an equivalent standard. This regulatory advantage can translate into competitiveness: those who adopt these references first will have an edge in training, selection, and certification of AI competencies.

Who it is for

The standard has a broad audience, with concrete and distinct uses.

Companies can use it to structure AI teams with better-defined roles and verifiable competencies, a practical advantage especially for those building internal AI functions for the first time.

Public administrations find an operational reference for implementing the Italian Strategy for Artificial Intelligence 2024–2026 and the Triennial Plan for Information Technology.

Universities, ITS Academies, and training institutions can use it as a basis for designing training paths more aligned with what the market is actually asking for.

Certification bodies can use it under Law 4/2013 on unregulated professions to develop AI skills certification schemes.

ICT professionals who want to qualify or reskill toward AI find in the standard a clear reference to guide their path.

What is still missing

It is worth saying plainly: UNI 11621-8:2026 is a voluntary standard. It does not oblige anyone, and it does not sanction anyone. The accreditation system for individual certification is not yet operational.

The standard lays the foundations. The supply chain — exams, recognized certificates, periodic updates — is the next step, and its activation depends on how many certification bodies decide to adopt it and how quickly companies and public administrations begin to reference it in selection processes and procurement specifications. It is not a limitation of the standard itself. It is the nature of every voluntary standard: its real weight is built through adoption.

A not-so-obvious starting point

In a market where AI roles and titles have multiplied without a shared reference, having for the first time an official document that establishes what it really means to do certain jobs is already something. The standard establishes a common vocabulary. It creates the conditions for a gradual professionalization. And it positions Italy as a European benchmark at a time when no other country has yet published an equivalent standard.

Antonio Savarese, management engineer and Head of Smart Services at Enel Italia, summed it up like this: “In a market flooded with ambiguous definitions, April 30, 2026 marks a historic turning point. The publication of the UNI 11621-8:2026 standard brings order to the chaos, officially defining the professional role profiles for AI. This is not just bureaucracy: it is the compass that will guide companies and professionals in the years ahead.”

At a time when AI adoption is accelerating and the quality of competencies has become an industrial issue before a technological one, having a shared compass is a starting point that, until a few weeks ago, did not exist.

Sources: