AI readiness to implementation for Argentina’s education systems

UNICEF Argentina, PAIDEIA and the Secretaría de Educación · Buenos Aires · October 2026

Turning AI interest into decisions someone owns.

A three-hour, hands-on workshop for members of the Red Federal PAIDEIA, the federal network of education officials and technical teams working on AI. The aim was to help them analyse opportunities, make informed decisions and guide AI implementation with learning at the centre.

My role: Facilitator, as Director of Product and Innovation at Cambridge. I designed and led the session, including the slide deck, the participant workbook and the four activities.

16 provincesOfficials and technical teams from across the country
3 hoursFour practical activities in one session
Six pillarsA shared framework for what else must be in place
One action eachWith an owner, the evidence needed and a review date
Niall McNulty with two colleagues in front of a UNICEF banner at the PAIDEIA workshop in Buenos Aires.

The brief

Practical cases, not a technology tour.

UNICEF Argentina, PAIDEIA and the Secretaría de Educación brought together members of the Red Federal PAIDEIA from Córdoba, La Pampa, Neuquén, Chubut, Salta, Corrientes, Mendoza, CABA, Provincia de Buenos Aires, Tierra del Fuego, Entre Ríos, La Rioja, San Juan, Río Negro, Misiones and Tucumán. The focus was practical cases of AI in education management.

The people in the room make or shape decisions about tools, data and policy. They do not need to build AI. They need to judge it: to recognise what kind of tool a proposal describes, know what it can get wrong, and ask for the evidence that it helps.

Design problem

How could one short session move a mixed group of national and provincial staff from a general interest in AI to a concrete next step they could take back to their own teams?

My contribution

Designing for a decision, not a presentation

Every part of the session pointed at the same output. Participants arrived with one AI priority or question and left with one feasible action, an owner, the evidence needed and a review date.

Start with the basics that matter for decisions

We separated three kinds of tools: generative AI, prediction and routine automation. Routine automation follows set rules and is not AI, though it is sometimes sold as AI. Each kind fails in different ways, so each needs different checks. Four questions ran through every example: what is it for, what is the benefit, what should a person check, and what evidence would show it helps?

Use a shared framework for readiness

The Six Pillars Framework from UNESCO, ITU, UNICEF and GPE (2024) gave the group a common way to ask what else must change before an AI idea can work: coordination and leadership, content and solutions, connectivity and infrastructure, data and evidence, capacity and culture, and cost and sustainability.

Work through a realistic case

Tables worked on Sierra Verde, a fictional province weighing up an AI tool proposal. The case set out the proposal, the data, the users, the resources and limits, and what the ministry did not yet know. It gave everyone the same problem to test ideas on before turning to their own.

Ground governance in local law and real roles

The policy section linked principles to Argentina’s own rules, including the Personal Data Protection Law (Law 25.326) and the Law on the Protection of the Rights of Children and Adolescents (Law 26.061). It also asked who decides what at national, provincial and school level.

How it ran

Four activities, each building on the last

1. Sort and checkParticipants sorted everyday examples, from a chatbot that drafts practice questions to a system that gives students a risk level for leaving school, into generative AI, prediction, routine automation or not sure. Pairs and tables then compared answers and reasons.
2. Map the dependenciesUsing the six pillars, tables mapped what the Sierra Verde proposal relied on and what was most likely to block it.
3. Decide and assignTables turned principles such as human oversight into a named person, a control and the evidence it works, and agreed whether each was decided nationally, provincially or by schools.
4. Plan your next stepEach person completed an action canvas, using the case or a real priority, and left with one action, an owner, the evidence needed and a review date.

Each activity followed the same rhythm: a short step on your own, then in pairs, then at your table. A printed workbook and A3 sheets kept the thinking on paper and gave people something to take home.

Provincial education officials working in table groups on printed workbook activities during the PAIDEIA workshop in Buenos Aires.

UNICEF Argentina shared that participants found the session very relevant to their work. Read UNICEF Argentina’s post on LinkedIn.

What I took from the work

Three design lessons

  • Officials need to judge AI, not build it. Naming the kind of tool and its likely failures gave people a firmer footing than any amount of technical detail.
  • Readiness is about everything around the tool. The six pillars helped tables see that capacity, data or cost is often the real blocker, not the AI itself.
  • A principle only works when someone acts on it. If you cannot name a person, a control and the evidence, the principle is only a wish.

Attribution: The workshop was convened by UNICEF Argentina with PAIDEIA and the Secretaría de Educación. This case study focuses on my contribution as designer and facilitator and does not imply sole ownership of the programme or its outcomes.

For the thinking behind this approach, see AI Readiness for Schools: A Practical Framework, Sovereign AI for Education Systems: A Practical Guide for Ministries and How to Design AI Professional Development That Changes Practice. Related writing: AI in education: practical lessons from implementation and The question is not which AI tools to buy.

Sources and further reading

  • UNICEF Argentina, LinkedIn post on the PAIDEIA workshop (October 2026).
  • UNESCO, ITU, UNICEF and GPE (2024), Six pillars for the digital transformation of education: A common framework.

Explore more work

See other programmes spanning system leadership, publishing and professional learning.

Back to selected work