Only 4.1% of Spain's industrial SMEs use AI: how to start

What the IndesIA Barometer 2026 says

IndesIA is an association that promotes the use of data and artificial intelligence in Spanish industry. This year's barometer, built on a sample of 131,554 companies, puts AI use at 4.1% of industrial SMEs, up from 2.9% in 2025 and 2.1% in 2024, as reported by El Español.

Size matters. AI is present in 6.5% of medium-sized companies (50 to 250 employees), 5.1% of small ones (10 to 50) and 3.6% of micro-enterprises. By region, Madrid leads with 6.3%, followed by Catalonia with 4.9% and the Basque Country with 4.2%, according to Conecta Industria.

At the forum where the barometer was presented, Lieven Van der Veken of QuantumBlack, McKinsey's AI arm, shared another figure, this one about organisations in general: 89% have adopted AI in at least one function, but only 6% see a real impact on their results, Conecta Industria reports. Having a tool is not the same as changing how the work gets done. The barometer also shows where the technology is heading: generative AI and decision automation are gaining ground over classic machine learning solutions, La Ecuación Digital notes.

An ordinary Tuesday in a twenty-person company

Imagine you run a machining company with twenty people. Orders come in by email, by phone and over WhatsApp, and someone copies them into a spreadsheet. Quotes are always prepared by the same person, because nobody else knows what each part costs, and that person spends half the day on the shop floor.

A customer asks for a price on Monday and gets it on Wednesday. Sometimes, by then, they have already bought elsewhere. You have read about artificial intelligence and assumed it was for companies with a data department. Then the phone rang and you left it for another day.

Why doesn't AI reach small companies?

The scene is not any one company, but many will recognise it. In a small business the day-to-day swallows everything. Nobody has a free afternoon to think about how the company should run two years from now, because this week's order has to go out.

On top of that come three beliefs that hold people back: that AI demands a huge investment, that you need perfect data, and that everything has to change at once. None of the three is true, but any one of them is enough to put the decision off. The barometer points the same way: the smaller the company, the less AI is present.

What staying the same costs

Staying the same is not free. It is hours every week copying data from one place to another. It is quotes that arrive late. It is knowledge that lives in one person's head and goes on holiday with them.

And there is a cost you do not see until it is too late: the company next door that did take the step and answers in an hour what takes you two days. Today it is 4 in every 100. Two years ago it was 2.

Where does an SME start with AI?

Not with the technology, but with a problem. The tasks that suit it best are repetitive and work with text or documents: reading orders, preparing quotes, answering the same customer questions, searching technical documentation, writing reports. The order that works best is this:

• Pick a single process that repeats every week and hurts: quotes, orders, customer service, documentation.
• Look at the data you already have. Emails, spreadsheets and your management software are usually enough to begin.
• Build a first version that solves only that, together with the people who do that work today.
• Measure the result in something you understand: hours saved, response time, errors.
• Once it works, move on to the next process.

In the machining company, that first version could read the orders that arrive by email, draft the quote from previous ones and leave it ready for a person to review and send. The decision still belongs to a person. What disappears is the copying and the searching.

How we help at digizone

At digizone we design and build custom software, and we take care of implementing artificial intelligence in companies: understanding the process, deciding what is worth automating and building it to fit the tools you already use. That is the difference between having a tool and feeling the result.

A recent example is Orlapp, a platform for graduation class composites. It started from a process that relied on in-person photo shoots, emails and a great deal of manual work. Today every student gets an AI-generated portrait from a selfie, and what used to be in-person and manual is a digital flow you reach from your phone. It is a different sector, but the starting point was the same as any SME's: a manual process someone decided to rethink.

Start with a first version

We call that first version an MVP, a minimum viable product: the smallest thing that already solves a real problem. It is not a mock-up to show around. It is something you use from day one and that grows with what you learn.

What matters is to start with something and scale it from there. The 96% that still does not use AI does not need a five-year plan. It needs one process solved.

If you would like to see how it would work for you, our free workshop, in Spanish, walks step by step through building an MVP in four weeks.