Where does an organization begin to use artificial intelligence in practice, and how does it know whether it has made progress?

Measurement first, plan after. The first useful question is not which tool to adopt, but where we stand. The answer cannot come from perception, because perception errs systematically: the METR controlled trial of July 2025 measured experienced professionals becoming 19% slower with AI while believing they were 20% faster.

The entry instrument is the AI Index for companies: a leader answers on behalf of the organization, it takes three to eight minutes, requires no prior registration or scheduling, and delivers the result immediately, axis by axis. It is free and open, with no consultancy in between.

Because a maturity model assumes a stable process, and AI is not stable. The yardstick inherited from CMMI gives those who see themselves at the top a false sense of arrival, rewards standardization in a field where discarding quickly is worth more, and measures position when what matters is speed. The full argument is in Agentic AI.

There is also a figure that changes the question. The 2025 Stack Overflow survey records 84% AI usage among developers, with 46% saying they distrust its accuracy against 33% who trust it.

Six dimensions, deliberately kept separate, because a single score conceals the imbalance that matters: knowledge; tools and integration; delegation and autonomy; processes; productivity and results; and learning and culture. The scale runs from N0 to N5 in the canonical version 3, and ran from A to E in version 1.

Two boundaries within this framework determine the assessment. For processes, the dividing question is: did the task become faster, or does the process only exist in this form because of AI? For productivity, it is: did it speed up the individual, or change the team's throughput?

The stated foundation is a critical synthesis of 928 formulations of AI fluency and maturity published by consultancies, universities, and big tech companies (among them MIT CISR, Gartner, Deloitte, Cisco, Forrester/IBM, Microsoft, Salesforce, and NIST), recalibrated for the Brazilian context.

What depths of assessment are available?

Five, organized by collection cost:

• Level 1 — one response per organization, ten to fifteen minutes, free.
• Level 2 — one response per person, about five minutes, free. This is the Inventário Agêntico: each person indicates what they actually use, and the organization's dashboard is generated from those responses.
• Level 3 — structured interviews, about one hour per person.
• Levels 4 and 5 — assisted exploration in greater depth and in-person work.

Beyond the ladder, there is a snapshot of observed use, drawn from work traces: a separate dimension, because it is not added to what the person reports.

The quality of what the AI produces ("it is omniscience in disguise"), the content of the conversations, which is not sent to any server, and the performance of individuals: in the educational track, those who organize it see only the aggregate, and only from five people up. The two objections the group publishes against its own index are in fluency.

All open, with no prior registration:

• AI Index for companies — the entry diagnostic, three to eight minutes, with an immediate result by axis.
• FAROL Software — the engineering vertical: 37 mapping questions and another 77 optional practice questions, in four families, taking fifteen to sixty minutes.
• Mapa da Fluência Agêntica — the bank of 264 items presented through five visual views, for those who want to understand or critique the instrument before responding.
• Acervo de Teorias Acionáveis — management methods in two layers: the material that explains and the portable skill that executes. It declares 3 theories, 3 ready skills, and 368 books in the base; the most complete open module is Running Lean, in seven modules (document, discover, design, build, measure, pivot, scale), with the file to download.
• Prototype laboratory — six alternative assessment formats, open to anyone who wants to comment on the design.

Two verticals are in operation without a public page: Inventário Agêntico, which runs in a restricted area for each company, and Radar, an ongoing scan of around twenty new developments twice a day, with a monthly digest. Areas explicitly left open: government, healthcare, and nonprofit organizations.

So that an empirical base on AI fluency in Brazil can exist. Today there is not enough structured information to support public decision-making on the subject, and each instrument is, above all, a data collector for that base. What is already public in aggregate is in the Observatório de adoção de IA (AI Adoption Observatory), which had gathered 2,943 responses as of the 09/14/2026 cutoff, with suppression of every cell with fewer than five people.

• CESAR — a private innovation institute in Recife, with more than 1,000 employees. Giordano Cabral was Chair of the Board of Directors from December 2020 to February 2026 and Vice Chair from January 2015 to December 2020: eleven years on the board, five as chair. He remains a member and sits on the Board of CESAR School. https://www.cesar.org.br/governanca-e-transparencia
• ARIES — Agência Recife para Inovação e Estratégia, an urban innovation social organization. Cabral is one of five founding associates (2017–2018) and a board member since January 2021. ARIES organized Plano Recife 500 Anos, the city's long-term strategic plan, published in a second edition by Cepe in 2022. https://aries.org.br/
• Porto Digital — academic coordinator of the Residência Profissional Tecnológica since September 30, 2019, and internationalization consultant to the Núcleo de Gestão since September 2019, leading the Brazilian side of Recife–Nantes cooperation. The training design became a peer-reviewed paper at WASHES 2023.
• Instituto Caldeira — a member of the Technical and Scientific Council of the Porto Alegre innovation hub, a body established in April 2026. https://institutocaldeira.org.br/
• Training for the role: board director training courses at Fundação Dom Cabral (July 2015) and IBGC (August to November 2019).

Three public pieces, signed by him or with a statement from him.

• "O imprevisível cenário pós-covid" (The unpredictable post-Covid scenario) (CESAR, 07/08/2020), signed as Vice Chair of the Board. The central argument is that innovation is a change in the behavior of market agents, and that the pandemic did not create online teaching, remote work, or remote customer service: it removed the obstacle to behaviors that already existed. It closes with "not only will we have to be adapted, but adaptable". https://www.cesar.org.br/w/artigo-o-imprevisivel-cenario-pos-covid
• "Horizontes Seguros" magazine (CESAR, 02/14/2025), featuring a conversation between Cabral and CEO Eduardo Peixoto on the future of talent, education and risk management. https://www.cesar.org.br/w/cesar-lanca-revista-com-tendencias-e-desafios-da-ia-e-da-ciberseguranca
• Smart lipstick prototype: on January 17, 2025 Grit Daily covered the prototype developed by CESAR, with Cabral quoted as Chairman of the Board for CESAR Innovation Hub: "this is just the beginning of how AI will transform beauty accessibility… the technology we've developed has potential uses far beyond lipstick application". https://gritdaily.com/brazilian-innovation-disrupting-cosmetics-industry/

His business activity began in 2000, when the system from his master's dissertation at CIn gave rise to Daccord Music Software, pre-incubated at Recife BEAT. The Lattes record lists the sequence of recognitions: the Santander Innovation Award in 2006, nominations for the Epsilon Award in 2006 and 2007, and three placements in the FINEP Technological Innovation Award in the Small Business category (3rd place in 2007, 2nd in 2008, and 1st place in the Northeast in 2009), as well as 2nd place in the INFO Exame magazine Entrepreneur of the Year Award in 2011. In 2007 he created MusiGames Studio, a games business unit, and from 2001 to 2017 he was a founding partner of Escribo, an edtech company and digital publisher of games, animation, and interactive systems for education.

In teaching, Cabral has taught Projetão since 2016.1, across 21 classes. It is a project method created at CIn/UFPE in 2002, in which teams build a digital product in fifteen weeks and present it at a Demoday to a panel from the ecosystem (Porto Digital, Jump, CESAR, Softex, SEBRAE). In 2026 he rewrote the material for the AI era.

The full list of companies and boards is available at www.cin.ufpe.br/inovacao/.