Frequently asked questions about AI in business

No. Today there are AI tools available from 20 to 40 euros per month per user, with no technical skills required. 58% of small American businesses already use generative AI, up from 23% just two years earlier.

The question is not the size of your business, but whether you have a concrete problem that AI can solve better than the alternatives. A professional who spends 8 hours a month on repetitive tasks can reclaim that time with a tool costing 30 euros a month. The return is immediate and measurable.

Costs vary enormously depending on the type of project. Ready-to-use tools (ChatGPT, Claude, Gemini in their professional versions) cost from 20 to 40 euros per month per user. Specialised tools (transcription, customer service, email automation) range from 10 to 300 euros per month. At this level, a small professional practice can start with an annual investment of less than 500 euros.

Structured projects with integrations between business systems start at 5,000 euros. Enterprise solutions for supply chain or production exceed 50,000 euros and require months of implementation. The book analyses concrete solutions organised by three budget ranges: under 5,000 euros, between 5,000 and 50,000, and over 50,000.

Not because of the technology. According to an MIT study, the vast majority of projects fail due to avoidable errors in the planning and implementation phases. The most common: looking for a technological solution before clearly defining the problem, having unrealistic expectations about results, underestimating the importance of data quality, and failing to plan a process for human validation of outputs.

A concrete example: a healthcare organisation invested in an AI system to optimise appointment scheduling. The system failed because the real problem was not technological: the scheduling process itself was disorganised, with unclear rules and inconsistent criteria. AI cannot solve an organisational problem.

For generative AI tools, no. Anyone can open a browser and start using ChatGPT or similar tools without writing code, without complex configuration and without technical skills. 53% of large Italian companies have already purchased generative AI tool licences.

For more complex projects requiring integrations with business systems, technical skills are needed, but they can be purchased externally. The skill that truly matters is strategic: understanding where AI creates concrete value and where it is an unnecessary complication. This is precisely the competence the book helps develop.

AI makes sense when several conditions are met simultaneously: the problem is costly in time or money and recurs regularly; you have sufficient quality data or you use pre-trained tools; you can tolerate an error margin of 5-15% or you provide human oversight; you have already tried simpler solutions and they were not enough; the value generated clearly exceeds the costs of implementation and maintenance.

If even one of these conditions is missing, it is worth pausing and reconsidering. The book includes a self-assessment toolkit to evaluate your situation in a structured way.

The available data does not support catastrophic scenarios. The World Economic Forum estimates that by 2030 AI will create 170 million new jobs and eliminate 92 million existing ones, a net positive of 78 million. Similar predictions have accompanied every transformative technology of the past two centuries, from mechanical looms to computers.

In practice, AI transforms jobs more than it eliminates them: it automates repetitive tasks and frees up time for activities that require expertise, relationships and judgement. Many SMEs thrive without AI and will continue to do so. AI is a useful tool in certain contexts, not a necessary condition for survival.

Start with a problem, not with the technology. Identify a task that costs you time or money, recurs regularly, and is currently handled inefficiently. Before thinking about AI, try simpler solutions: a well-structured spreadsheet, simple rule-based automation, better process organisation.

If simpler solutions are not enough, then evaluate AI with realistic expectations: define measurable success criteria, start with a contained project and verify results before extending the investment. The book presents a structured five-phase method for this journey.