When an AI system flags a change in an older person’s routine, a care professional still has to decide whether to call or visit. When a model suggests irrigating a crop, a grower has to judge whether the recommendation fits the conditions on the ground. Both examples raise a question that ran through the series: what do professionals need to know to act on an AI recommendation?
From 23 October to 10 November 2026, AI-SECRETT explored decisions like these in seven webinars during the European AI Innovation Month. Drawing on the project’s sector research, the series examined how the digital, green and social transitions intersect in professional practice. Each session paired a sector presentation with a Fast-Sight® exercise, inviting participants to imagine how their work might change by 2030 and identify the skills they would need.
23 October · Letizia Sgalambro and Peter Voelk, TÜV Thüringen Italia, and Matteo Sirizzotti, Cognito
Letizia Sgalambro presented a virtual village for older people living alone with limited support networks. A device at home connects residents to calls, reminders and activities. AI could personalise activities or flag changes in routine for care professionals to assess.
Matteo Sirizzotti explored how game design can help people practise safely and investigate accessibility. In one example, students explored virtual reconstructions of heritage sites and identified barriers such as uneven floors and missing tactile signs. Peter Voelk took the discussion into certification, showing how AI could help auditors review evidence and identify emerging risks. One option for sensitive documents was to use an approved model offline, with the auditor responsible for interpreting and challenging its outputs.
27 October · Elena Silvestrini and Marta Anducas, Lisbon Council
Elena Silvestrini set the discussion against Europe’s housing crisis and the EU response. According to Eurostat, house prices in the EU rose by 60.5% between 2010 and the second quarter of 2025. The session explored how digital tools could support housing policy by helping professionals identify areas under housing pressure and compare renovation options.
Marta Anducas showed how these ideas translate into practice. London’s Office of Technology and Innovation identified opportunities in housing maintenance and resource planning. The Syn.ikia pilot in the Netherlands explored how digital twins could help make better use of surplus solar energy. In Canada, Malahat Nation illustrated how generative design can give a community more options to discuss and refine. The session connected these examples to a practical test for renovation programmes: can they improve energy performance while keeping homes affordable for the people who live in them?
29 October · Chema Segovia, Culturalink
A steel mill, a car plant and a furniture workshop create value in different ways. Chema Segovia used those examples to explain why AI adoption depends on the production context. A digital twin can help test a change before a manufacturer invests in it. Generative design can suggest a lighter part or one that is easier to disassemble. The feasibility of either proposal depends on people who understand the materials, machinery and production conditions.
That knowledge also matters when something goes wrong. In the Fast-Sight® poll, eight of 16 respondents said the hardest decision was determining when workers should be able to question, stop or override an AI recommendation.
2 November · Pau Rausell Köster, Econcult, Universitat de València
Using the Global Innovation Index 2026, Pau Rausell Köster compared eleven European economies and showed how strengths in design, cultural production and digital activity can coexist in very different combinations. The creative outputs pillar accounts for a quarter of the index. Italy, for example, ranked fourth in industrial designs and 66th in mobile app creation in the comparison presented.
The discussion then turned to how generative AI and platforms shape the way creative work is produced, distributed and paid for. In the exercise, the most selected challenge was handling copyright, consent and fair remuneration for creative work used by AI. Participants gave the highest training priority to using AI critically and evaluating its outputs, followed by developing business models and managing intellectual property and finance.
4 November · Andreas Almpanis, JOIST Innovation Park
Andreas Almpanis used an irrigation decision to show what AI can contribute. A system can combine forecasts, soil and crop observations, historical patterns and sensor or satellite data to suggest when to act. A grower then weighs the recommendation against local conditions, practical experience and costs.
The session also asked whose circumstances are represented in the data and who can benefit from them. Smaller or less connected producers may be harder to capture in datasets built around sensors and standardised systems. In the Fast-Sight® exercise, six of ten respondents selected access to AI without dependence on a single supplier as the hardest decision for farmers, cooperatives and small businesses. The highest-scoring training priority was framing real problems and combining AI with local and expert knowledge.
6 November · Catalina Bohórquez, CINK Venturing
Catalina Bohórquez examined what happens when AI becomes part of the infrastructure used to forecast electricity demand, balance grids and monitor environmental change. The presentation identified three concerns: dependence on a small number of providers, the energy, water and materials used by AI itself, and the need for transparent decisions that affected communities can help shape.
The research team reviewed 29 training programmes across 16 countries and pan-European consortia and found that none explicitly covered the obligations for high-risk AI systems in critical energy infrastructure. The session identified four areas for training: technical AI and data skills, ethics and governance, digital sustainability, and creative and systemic thinking. One participant made the operational implications concrete by proposing that staff should be able to run critical energy systems manually if an AI system fails.
10 November · Korina Kominou, INESC TEC
Korina Kominou showed how AI could help explore multimodal routes, anticipate infrastructure failures and coordinate electric vehicle charging with the grid. The session treated accessibility, affordability and job quality as criteria for judging those changes. A route can look inefficient in a dataset and still provide an essential connection for the people who use it.
In the Fast-Sight® poll, five of nine respondents said the hardest decision was how to align data, standards and investment across operators, public authorities and borders. When asked which skills to prioritise, participants gave the most points to assessing the environmental impact of mobility choices, followed by assessing accessibility and affordability for different users and territories. In their written responses, participants also proposed protecting routes at risk of withdrawal and linking transport planning with housing and employment.
Across the seven webinars, a common question kept coming up: how should professionals decide what to do with an AI recommendation? The answer depends on the work. A grower may need to adapt a forecast to local conditions, an auditor may need to challenge the evidence behind a recommendation, and a transport planner may need to consider who still depends on a route that looks inefficient in the data. In each case, professional judgement helps determine whether a digital change also delivers environmental and social benefits.
The Fast-Sight® exercises, led by Inspiring Futures Europe, gave participants a chance to identify the skills they would need in those situations. Their responses highlighted practical priorities for AI-SECRETT’s training: evaluating evidence, knowing when to intervene, weighing environmental and social impacts, and involving workers and communities in decisions that affect them. These priorities can inform training built around real workplace decisions, where professionals have to explain why a recommendation is worth following, adapting or setting aside.
Thank you to the speakers and partners who shared their research and examples, and to the participants who brought their own experience to the discussion. Their contributions helped connect the project’s sector research with the decisions professionals face in their day-to-day work.