Ask any company whether it uses artificial intelligence and the answer is often "not yet". Ask the people who work there and you hear a different story. In Microsoft and LinkedIn's 2024 Work Trend Index, a survey of 31,000 people in 31 countries, 75% of knowledge workers said they already used generative AI at work, and 78% of those users were bringing their own tools, without guidance from their company.
This is "shadow AI": useful, spontaneous and completely unmanaged. Quotes drafted on personal accounts, customer data pasted into free chatbots, results that nobody measures. The risk is twofold. Company data leaves your control, and the real potential stays unused because every person reinvents the wheel alone.
The AI Evolution Program, Vivig's AI training path for companies, was designed for exactly this situation. Here is how it works in practice, and what we have learned building it.
Why a programme and not a course
A one-day AI course is pleasant and quickly forgotten. People go back to their desks, the old habits return and nothing changes in the processes. That is why the programme is built as a change project of 12 weeks, with one meeting a week at the company and analysis and development work in between. About 80% of the time is spent on the company's real cases: its quotes, its invoices, its customer requests.
The goal is not "knowing AI". It is ending with tools in production, trained people and numbers that show the difference.
The five phases
- Diagnosis (weeks 1–3). Interviews with the people who do the work, a map of the processes, the Top 10 use cases ranked by impact and the starting KPIs: hours spent, response times, errors.
- Foundations (weeks 4–5). A secure company AI environment, set up on Claude for Business with users, projects per department and a shared knowledge base. Alongside it, an AI usage policy and a prompt library for each role.
- Skills (weeks 6–8). Training for everyone and hands-on labs for each department, plus a network of internal AI Champions who keep the method alive after the programme ends.
- Automation (weeks 9–11). Internal apps, company Skills, scheduled tasks, agents and integrations with the tools the company already uses.
- Measure and ROI (week 12). A demo day, a before-and-after report on the KPIs and a 12-month AI roadmap.
What changes, team by team
- Sales: from a customer request to a draft quote in minutes, with company price lists and terms already applied.
- Administration: invoices and documents read and summarised, payment reminders and monthly reports prepared automatically.
- Customer care: consistent, fast replies based on the company's FAQs and procedures, with requests routed to the right person.
- Management: data analysis, meeting minutes and a weekly report that arrives on its own every Monday.
In every case people keep the final word. AI prepares, the person checks and decides.
Compliance as a by-product, not an extra
Since February 2025, Article 4 of the EU AI Act requires companies that provide or use AI systems to take measures to support the AI literacy of their staff. The 2026 Digital Omnibus kept this obligation, while making clear that no specific level is mandated. The European Commission's Q&A is practical on two points: there is no need for a certificate, and simply asking staff to read the instructions for use is unlikely to be enough. An internal record of training and initiatives is the sensible way to show what has been done.
In Italy, Law 132/2025, in force since 10 October 2025, adds that employers must inform workers when AI is used in the employment relationship, and that professionals must tell clients clearly when they use AI systems in their work.
A well-run adoption project covers these points naturally: the usage policy says which tools and data are allowed, the training is documented with dates, materials and attendance, and the process map shows where AI is actually used. Compliance becomes a by-product of good work rather than a separate paperwork exercise.
Three lessons for any company starting with AI
- Start from processes, not from tools. The right question is not "which AI should we buy?" but "where do we lose the most hours?".
- Measure before you start. Without starting KPIs, any result is just an impression.
- Give the method an owner inside the company. AI Champions in each department make the difference between a project that lasts and one that fades in three months.
How Vivig can help
Vivig Consulting builds AI into real products, such as the SafeBloom, NettoChef and SafeGuard apps you can find on our projects page, and brings the same practical approach to company teams. The AI Evolution Program comes in three formats, from small businesses to companies with several sites. If you would like to see what it would look like in your company, write to us.
This article is general information, not legal advice.