Small Polish firms start AI with marketing. How to do it without a designer

How many Polish companies actually use AI

Eurostat has the most comparable figures, and they aren’t flattering. In 2025 only 8.36% of Polish firms with ten or more staff used any AI at all. The EU average was 19.95%, and Poland sat near the bottom of the table. Split by size, it looks worse still. Among small firms of 10 to 49 people the share was 5.79%, against 17% across the EU. Medium-sized firms did better at 14.77%, but the EU figure there was 30.36%.

You’ll find higher numbers elsewhere. A report published in April 2026 by PARP, Poland’s enterprise development agency, together with Jagiellonian University puts the figure at 23%. That survey, though, only looked at twelve sectors where AI has the most potential, not at the whole economy. When the Polish Economic Institute lined several studies up side by side in September 2025, it got anything between 5.9 and 16%, depending on how the question was asked. Whichever number you pick, the story is the same: Polish companies, and small ones in particular, are only getting started.

Marketing goes first

Look at what the early adopters actually do with it and a pattern jumps out. Eurostat’s data says 65% of small Polish firms that use AI put it to work in marketing or sales. Across the EU that’s just 34%. PARP saw the same thing from another angle: the most common use is content creation and creative support (texts, images, code, presentations), named by 33% of companies that already use AI. When a small Polish company picks up AI, it’s mostly to get content out faster.

There is a gap here, though. Only 2.31% of Polish SMEs use tools that generate images, video or audio, against 9% in the EU. So text is increasingly made with AI, while the visual side, the one customers recognise a brand by fastest, lags behind. On top of that, according to PARP, 64% of companies using AI rely mainly on free, general-purpose tools.

The Polish Economic Institute adds one more observation. Its interviews with companies showed that AI often arrives bottom-up, from employees, rather than from company strategy. One interviewee, responsible for content at a medium-sized advertising firm, talked about spending less time on graphics, video and presentations. We would add one question: in a setup like that, who makes sure all those materials look right together?

NAOKO: AI in design, not instead of it

There aren’t many well-documented Polish cases from smaller firms yet. We found two worth telling. The first is NAOKO, a fashion label set up in 2016, which sold over 50 million zloty of clothing online in 2022. A year later it started a project called NAOKO.AI. Customers co-created a collection through surveys, workshops and video responses, and AI processed their answers, proposed the shape of the collection, moodboards and product visualisations. AI-generated figures were meant to show how a piece looks on different body types and sizes.

As the owners told Fashion Biznes, the company processed more than 700,000 responses, and the collection kept the designs that at least 80% of the community liked. One important detail: everything AI produced still went through the team, and the designers had more work on this project than usual.

„Bez AI nie zrobilibyśmy tego.”

“We couldn’t have done this without AI” (translated by CAASUAL).

For us this is a good example of bringing AI into the design process sensibly. It didn’t replace the designers, and the owners themselves point out that this way of working isn’t cheaper than the traditional one. What it did was let the company do something it couldn’t have done otherwise: design a collection together with tens of thousands of customers.

Lisek.App: AI graphics, but in the brand’s style

The second example is smaller and more everyday. As the Polish media site Wirtualnemedia.pl reported in November 2023, the marketing team at the start-up Lisek.App uses Midjourney to produce graphics from a specific brief, for the brand’s own channels and for ad campaigns. That’s not the end of the process. The generated images go to a graphic designer and an art director, who adapt them to the brand’s visual language and add its signature elements.

According to the company, this cut the time to prepare one graphic from several hours to about a dozen minutes, and graphics made with AI’s help made up around half of all marketing materials at the time. The company also said it no longer needed outside creative agencies. These are its own claims, but the shape of the process is worth noting: AI makes the first version, and people who know the brand decide whether the graphic looks like it.

And a warning from December 2025

The opposite example is Polish too. In December 2025, the brand Dafi posted a spot that, according to media reports, was entirely AI-generated and quickly disappeared after a wave of criticism. We tell that story in our post on AI slop and your brand. The lesson matters even more for small companies, because one failed ad weighs more there than at a global corporation.

Why doing it alone with free tools isn’t enough

Polish companies that have considered AI name lack of expertise as the main barrier. Among SMEs, 63% say so, according to Eurostat. The EFL AI Barometer for the first half of 2026, reported by Bank.pl, shows that only 7% of SMEs rate their AI readiness as high or very high, and only 3% of micro firms. At the same time, 97% develop AI on their own. 80% of medium-sized firms work with independent experts, compared with 33% of small firms and 11% of micro firms.

Commenting on the survey, EFL warns that with readiness this low, going it alone easily leads to fragmented rollouts without a plan, safeguards or a way to measure results. In marketing and content this usually means one thing: everyone makes materials their own way, and the brand slowly blurs.

Where to start if you have no designer

A company without a design department can start sensibly and without a big budget. We suggest this order:

1. Write the brand down in a way AI can use: logo, colours, fonts, how you talk to customers, plus a handful of past materials you were happy with.
2. Get a few templates made for the things you produce most often. For most small firms that’s a post, a banner, an offer and a job ad.
3. Choose one repetitive job to start with. Weekly posts work well, and so do language versions of a catalogue. Time how long it takes now.
4. Agree who signs content off before it goes out, and what AI should never generate for you.

This is where we start with companies that have no in-house designers: with an AI-readable brand that lets marketing, sales or HR make consistent materials on their own. Bigger projects can then come to us, including through our CAAS subscription. How to describe a brand for AI step by step is covered in our post on brand guidelines for AI. If the company doesn’t yet have a logo, colours and fonts for AI to work with, the first step is a visual identity project.