Tytus Golas and the small thing that became the whole business

Before there was an AI agent, there was an IKEA catalogue. At fifteen, Tytus Golas imported catalogues from Germany and sold them on Allegro, Poland’s online marketplace. The object was ordinary. The little business depended on noticing that something available in one place had value in another. A teenager had found a gap between supply and demand, with furniture photographs doing the heavy lifting.

Years later, a similar act of noticing would shape Tidio. Golas and his co-founder, Marcin Wiktor, were developing Tidio Elements, a tool for adding website widgets. It offered forms, cookie notices and live chat. Customers particularly liked the chat. The founders separated it into a product of its own. The useful part of a larger idea became the business.

Today Golas is Tidio’s co-founder and CEO, based in Warsaw, with a company rooted in Szczecin and serving businesses internationally. His public working life has taken an unexpected turn. After spending years learning how to manage an expanding organisation, he is writing software again. The keyboard has returned to a job that seemed to be moving steadily toward meetings.

The widget gets its own company

Golas had been making applications before Tidio. One Facebook project, Ranking Znajomych, let users rate which friends they found more appealing and produced a ranking. It worked well enough for him to try repeating the success. The later attempts generally went less well. Wiktor, meanwhile, had been developing a website for book enthusiasts. Neither arrived at customer service by following a tidy career ladder.

Tidio’s own account dates its founding to 2013 in Szczecin. Its development involved an agency business and experiments with online tools before the chat product became the focus. That matters because the business people know now was selected from things its founders had already tried. It did not require them to correctly predict every subsequent use of artificial intelligence.

A chat window also gave them a modest, legible problem. A visitor has a question. A business needs to answer. The gap can cost a sale or leave somebody waiting. An online shop may have the ambitions of a department store and the staffing arrangements of a kitchen table. Its customers still expect somebody to be there.

Golas has kept small and medium-sized businesses central to the proposition. He has described them as a market with an opening for better technology, rather than a consolation prize after the corporate customers are taken. He also drew a distinction between Tidio’s smaller customers and the larger businesses served by LiveChat, while acknowledging that customers moved between the two. Competition, in this telling, leaves room for neighbours.

Learning to leave the room

The product could spread faster than its founder could learn every new version of his job. In 2019, Golas described the difference between managing five people, fifteen and forty. First came distribution difficulties, then recruitment difficulties, then the discovery that adding colleagues changes the machinery of the company. Hiring somebody does not automatically remove the work of organising everybody.

The InCredibles acceleration programme, where Tidio was a finalist in 2017, helped alter his approach. Mentors challenged the assumption that longer hours and constant firefighting were sufficient. Golas recalled learning that the founders could not remain involved in every process if the business was to grow quickly. Delegation was part of the product’s future, even though it never appeared in the chat window.

There is an awkward bargain here for anybody who enjoys making things. The company succeeds; the maker’s day fills with decisions about other people’s work. In his January 2025 conversation with Greg Albrecht, Golas said that coding had been the main effort when he started Tidio, and that he had loved it. As the company scaled, he stopped. Something was missing.

His account of leadership also includes asking for help. By 2024, he was describing the importance of recognising where somebody else knew more, and enjoying colleagues who were on the same wavelength. He compared business to a game he liked playing. That is a useful description of his appetite for another attempt: the attraction lies partly in having more to work out.

Money to keep building

Tidio entered 2022 with a business that already earned money. Its reported 2021 revenue was $8.2 million, with $1.6 million in net profit. Those figures belong to that year. They show why the subsequent investment was a choice about expansion, rather than simply evidence that an idea had persuaded someone to write a cheque.

In May 2022, the company raised a $25 million Series B led by PeakSpan Capital, with participation from Inovo Venture Partners and Rafał Brzoska, the founder of InPost. PeakSpan’s investment case emphasised ease of use, pricing and features suited to smaller ecommerce merchants. Its backing put money behind the customer group Golas had already chosen.

His ambition at the time was for one in ten online retailers worldwide to use Tidio by 2030. The date is an ambition, and the fraction describes the scale of the task. A chat widget may be small on a screen; distributing one to that many businesses calls for a much larger operation.

Golas has also stressed the Polish engineering work behind the international reach. In 2024 he described Tidio as having gone global once it focused on the live chat widget, with its technology continuously developed in Poland. The geography of software permits an interesting division of labour: the product travels while much of the building stays close to its original home.

Six years with a framework he stopped believing

Golas is willing to describe decisions he would reverse. In November 2024, he wrote about introducing objectives and key results, or OKRs, in 2017 after a conversation with an investment fund. The company had twenty people. He presented the framework with confidence, invoking the companies that used it. Engineering promptly asked how it was supposed to work for them.

He kept adjusting the approach for six years. What bothered him was the mismatch between recorded success and useful work: people making a difference could miss their targets, while others ticked the boxes. He came to prefer a simpler system with fewer measures. His conclusion is a judgement about his own experience, rather than a universal law of management.

There is some comedy in borrowing an admired company’s administrative habits before knowing whether they suit your own. A spreadsheet can acquire the authority of scripture remarkably quickly. Golas’s account is interesting because he includes his original enthusiasm. He had helped install the thing he later wanted removed.

His separate approach to product work draws a distinction between established services that must be dependable and uncertain experiments that need rapid feedback. Hosting, security and core features belong to the first group. An early test of an idea belongs to the second. The distinction gives people permission to move quickly without treating a paying customer’s everyday tools as a rough draft.

The CEO returns to the keyboard

AI offered Golas another way into the work he had missed. In January 2026, he described a Tuesday that included building an internal CRM, a prospecting agent and a year-in-review web app, alongside four hours of meetings. These were his own descriptions of what he made. The calendar still contained the CEO job; development had moved back into the gaps.

By May, he was describing a four-person team assigned to build a new product. He reported writing 56,309 lines of its code himself, while expressly cautioning that lines of code were a proxy for building speed. The number is revealing mainly as a statement of participation. He wanted to know the product through the act of making it.

“A coding founder is no longer a phase you grow out of.”

Tytus Golas, May 2026

That assertion has a personal history behind it. Earlier growth required him to step back from doing everything. The new tools have made hands-on work attractive again. The interesting question is how he combines that access to detail with the responsibility to let other people make decisions. His career has given him reasons to value both.

In July 2026, Golas described work on an AI assistant operating in Slack. The intended customers were ecommerce teams with little time to code or complete their existing tasks. Tidio’s current product family includes Hubi AI, a marketing coworker, alongside its customer service tools. The practical appeal is familiar: take something the founders can build and make it usable by somebody running a shop.

Who takes the difficult question?

For all his enthusiasm about automation, Golas has written about the people required to keep it useful. His January 2025 customer service framework proposed an AI trainer who reviews conversations and improves the knowledge behind them. Human agents, in his account, would concentrate on complicated cases. He reported that Tidio had automated 60 percent of its own support without job losses at that point.

That is a dated company claim, and it describes a division of work. Someone still has to decide whether an answer helped, maintain the information and handle the question that refuses to fit. The administrative chores have a habit of surviving the arrival of impressive technology. They simply acquire new job descriptions.

Golas’s public record now includes four appearances on Forbes’s 30 Under 30 lists, an EY Entrepreneur of the Year finalist place and a company working across customer service and AI tools. Yet the episodes that explain him are smaller: a catalogue worth selling elsewhere, a widget customers kept choosing, a framework he eventually abandoned, a keyboard he came back to.

Each involves paying attention to what happens after an idea meets other people. That gives this story its continuity. The fifteen-year-old seller and the CEO writing code are separated by a company, investors and years of management. Both are occupied with the same practical question: what can I make useful enough that somebody else will want it?