Bringing AI inside your company has gone, in a single year, from a niche experiment to an everyday tool for one Italian business in six. According to the Eurostat report on the use of AI in the European Union, 2026 edition, 16.4% of Italian enterprises with at least 10 employees use at least one AI technology. In the previous survey the figure was 8.2%: it has doubled.
The gap has not closed, though. The EU average is 20%. In Italy the wider gap is internal: large companies are moving fast, small and medium-sized ones much less. The reason cited most often is not cost. It is a lack of skills, named by roughly seven in ten companies that considered AI and then did not adopt it.
This guide covers the latest data on AI adoption in business, the departments where it pays off first, the AI Act deadlines that apply to anyone running chatbots and virtual assistants, and a five-step plan to get started without wasting budget.
Table of contents
- How many companies use AI in Italy and Europe
- Where AI inside your company is used
- Practical examples of AI inside your company
- Updated statistics on AI in companies
- The AI Act and deadlines for companies
- How to get started with AI inside your company
- Conclusion
- FAQ
How many companies use AI in Italy and Europe
The most recent picture comes from two sources that complement each other. Eurostat measures how many companies use AI. The Artificial Intelligence Observatory of Politecnico di Milano measures how much they spend. Together they show an Italian AI market worth 1.8 billion euros, up 50% on the previous year, with 46% of that spending already going to generative AI or hybrid projects. The other 54% is "classic" machine learning: forecasting, classification and analysis of structured data.
Growth is concentrated, however. 71% of large Italian companies have started at least one AI project, against 8% of SMEs. Eurostat shows the same pattern across Europe: companies with more than 250 employees have IT teams, dedicated budgets and vendors who follow them closely, while a 30-person business usually has nobody who can spend even half a day a week figuring out where to begin. That is why an SME's first project has to be small, measurable and manageable without an in-house data scientist. More on that below.
Where AI inside your company is used
AI almost always enters a company through one specific department, not through a plan that covers everything at once. Returns come first where there are many repetitive tasks and plenty of data already in digital form: customer service, marketing, human resources, logistics and finance.
Customer care
Customer care is the most common starting point because the problem is clear and easy to measure: too many identical requests, overloaded agents, customers left waiting. A virtual assistant built on generative AI reads the company knowledge base (manuals, contract terms, FAQs) and answers in natural language, at night and at weekends too. When a request is sensitive, it hands the conversation to a human agent together with the history, so the customer does not have to repeat everything.
In this department, chatbots and virtual assistants are used to:
- answer frequently asked questions within seconds, on the website, WhatsApp and the phone
- provide 24/7 support without night shifts
- route complex requests to the agent with the right skills
- use the customer's history to give answers in context
- read the sentiment of conversations and flag unhappy customers
These benefits apply across industries: in the utilities sector, for example, chatbots in the energy market already handle contract questions, billing requests and rate changes on their own. If you want to see the actual savings, we ran the numbers in how to reduce customer service costs with AI.
Marketing and sales
In marketing and sales, AI inside your company shifts the work from one-size-fits-all campaigns to messages built on each customer's data. AI combines CRM records, purchase history, browsing behaviour and social media to:
- build micro-segments of customers with similar needs
- predict what a customer will buy and when
- write or suggest personalised emails, ads and offers
- recommend products based on past purchases
- spot customers likely to churn before they leave
- adjust prices to demand (dynamic pricing)
A typical case: an online sports retailer notices that people who buy trail running shoes come back about six months later. A predictive model flags customers approaching that window, and a WhatsApp assistant offers them the new model in the size they already have on file.
HR AI inside your company
In human resources, AI takes over administrative work and leaves more time for interviews, training and managing people. The most common applications are:
- screening hundreds of CVs
- candidate chatbots that explain the hiring process
- automatic interview scheduling
- personalised onboarding for new hires
- an internal assistant that answers questions on leave, benefits and policies
- analysing employee sentiment from anonymous surveys
The internal assistant is often the easiest project to launch, because it works on documents HR already has: staff handbook, collective agreement, procedures. Be careful with CV screening, though: it is classed as high-risk under the AI Act, with specific obligations covered below.
Supply chain and logistics
In the supply chain, AI works on numbers: demand forecasts, stock levels, delivery times. The main uses are:
- forecasting inventory from historical data, seasonality and promotions
- reordering automatically when an item drops below a threshold
- tracking goods in real time with IoT sensors
- calculating the most efficient delivery routes
- predicting failures in machinery and vehicles (predictive maintenance)
- guiding warehouse robots for picking and sorting
Financial services
Banks, insurers and lenders have large amounts of data and heavily regulated processes, which makes finance the sector that has used AI the longest. Recurring applications include:
- detecting anomalous transactions and fraud
- monitoring transactions for anti-money laundering (AML)
- automating data entry, reconciliation and invoice processing with RPA and AI
- assessing creditworthiness (credit scoring)
- offering automated investment advice (robo-advisors)
- keeping up with regulatory requirements (RegTech)
Practical examples of AI inside your company
The same technologies are used very differently depending on company size and sector.
An SME usually starts with a chatbot on its website or WhatsApp that answers frequent questions and collects sales leads outside office hours. Take a window and door manufacturer with ten employees: every week it gets dozens of identical questions about installation times, tax deductions and quotes. A virtual assistant answers the first two on its own and, for the third, collects measurements, photos and contact details, which the sales rep finds the next morning already organised in the CRM. The owner has not hired anyone, and the sales rep no longer spends the first hour of the day answering the same emails. From there, many SMEs move on to analysing sales data to manage stock and campaigns better.
Large companies work on several fronts at once: predictive maintenance on plants, quality control with computer vision, energy optimisation, virtual assistants serving thousands of customers a day. The challenge here is connecting AI to existing systems (ERP, CRM, ticketing) and controlling who can see which data. To understand how an agent that works inside these systems is built, our guide to enterprise AI agents covers architecture, use cases and costs.
In healthcare, AI helps read diagnostic images (X-rays, CT scans) to detect conditions at an early stage, and adapts treatment plans to each patient's clinical data. On the administrative side, virtual assistants handle bookings, cancellations and reminders, taking pressure off booking centres.
Updated statistics on AI in companies
Worldwide, AI in business is now the norm. The economy chapter of the Stanford AI Index 2026 reports that 88% of surveyed organisations use AI and 70% use generative AI in at least one business function. Global corporate investment in AI reached $581.7 billion, of which $344.7 billion was private investment.
These figures mostly describe large organisations, many of them in the United States, so they should not be compared one-to-one with Italy's 16.4%, which counts every company with 10 or more employees. They do show where the market is heading: software vendors are building AI into the products companies already use, from CRM to email. For many SMEs, the first experience with AI will not be a dedicated project but a new feature inside a tool they already pay for.
For Italy, three figures are worth remembering: a 1.8 billion euro market, 71% of large companies with at least one project under way, 8% of SMEs. The distance between the last two numbers is where the real growth is.
The AI Act and deadlines for companies
The AI Act, the European regulation on artificial intelligence (Regulation (EU) 2024/1689), is already partly in force, and this year its timetable changed. The Digital Omnibus on AI (Regulation (EU) 2026/1744), in force since 27 July 2026, pushed back the obligations for high-risk systems. The new timetable for companies looks like this:
- from 2 August 2026, the transparency obligations of Article 50 apply: anyone using a chatbot or voice assistant must tell people they are talking to an AI;
- from 2 December 2027, obligations apply to the high-risk systems listed in Annex III, including those used to select staff or assess creditworthiness;
- from 2 August 2028, obligations apply to AI built into already regulated products, such as machinery and medical devices.
For anyone using AI in customer care, the deadline that matters has already passed. The bot must introduce itself as a bot from the first message. A single clear sentence is enough, for example "Hi, I'm the virtual assistant of…". If your chatbot currently uses a human name and photo, fix it now. Companies using AI for CV screening or credit scoring have more time, but they will have to document their data, human oversight and risk management, and it is easier to start collecting that documentation now than right before the deadline.
How to get started with AI inside your company
Many AI projects fail for the same reason: they start from the technology ("we want to use ChatGPT") instead of a problem. This is the five-step path we recommend.
- Pick a measurable problem. Look for a repetitive task whose cost you can already measure: customer service calls, time spent answering the same emails, order entry errors. Write down the starting figure.
- Train the people who will use the tool. Since a lack of skills is the biggest barrier to adoption, half a day of training for the team involved is worth more than any advanced feature.
- Choose a platform that matches your skills. Without in-house developers, go for a no-code platform that connects to the systems you already use. Before signing, work out the total cost: licences, integrations and maintenance. Our guide on how much a chatbot costs breaks it down.
- Run a 60 to 90-day pilot. One channel, one department, written goals. At the end, compare the result with the starting figure.
- Measure and expand. If the pilot works, extend it to other channels or departments. If it does not, find out why before switching vendor: the problem is often the knowledge base, not the model.
If you want an assistant that does more than answer and actually takes actions (opening a ticket, changing an order, booking an appointment), read our guide on how to create AI agents.
Conclusion
The first step for an SME is smaller than it looks. Take the twenty questions your customer service gets most often, check that the answers are written down somewhere (a document, a web page, a manual) and use them as the basis for a virtual assistant on a single channel. Within two or three months you will know how many requests it resolved on its own and how much time it freed up for the team.
The mistake to avoid is the opposite: launching three projects at once, across three departments, without a starting figure. The number of Italian companies using AI doubled in a year, but the gap between large companies and SMEs is still wide. Companies that start now with a concrete project still have time to close it.
FAQ
What is AI inside a company?
It is the use of software that learns from data to do tasks that used to need a person: answering customers, forecasting demand, reading documents, spotting fraud.
How many companies in Italy use artificial intelligence?
According to Eurostat, 16.4% of Italian companies with at least 10 employees use at least one AI technology. The EU average is 20%. Among large Italian companies, 71% have at least one AI project under way.
What are the benefits of AI for a business?
Less repetitive work, answers for customers at any hour, more accurate sales and stock forecasts, and lower operating costs. The concrete benefit depends on the department you start with.
How can a small business start using AI?
By tackling one measurable problem, such as customers' frequent questions. A no-code chatbot on your website or WhatsApp can go live in a few weeks without in-house developers. After two or three months, look at the results and decide whether to expand.
What does the AI Act require if I use a chatbot?
Since 2 August 2026 you have to tell people clearly that they are talking to an artificial intelligence. A short introduction at the start of the conversation is enough.
Will AI replace workers?
In most cases it changes jobs more than it removes them. Repetitive tasks move to AI, while people handle complex cases, customer relationships and checking what the machine does.
How do you measure the ROI of AI in your company?
Before you start, write down the baseline: cost per contact, average response time, hours spent on a task. After the pilot, measure the same figure again and compare the difference with the total cost of the project, including licences and maintenance.
