Welyo’s AI bet for growth

Polish customer service technology provider Welyo is betting on AI to return to growth after its revenue fell in 2025. Backed by MCI Capital, it also plans to expand across CEE while keeping people at the heart of customer service.

Błażej Rychlik, prezes Welyo
Błażej Rychlik took over Welyo during a period of integration between the two companies and business transformation. The results are already beginning to show. Photo: press materials/Welyo
Loading the Elevenlabs Text to Speech AudioNative Player...

Welyo, valued at PLN 161 million (EUR 36.5 million) by investment fund MCI, is undergoing a major transformation. After revenue fell in 2025, the company aims to return to growth with a new technology platform and a revised business model. It then plans to consolidate the market in Central and Eastern Europe.

It is difficult to spread your wings when operating in just one country, especially in technology. Welyo, a provider of customer experience (CX) software for managing customer interactions, knows this well. Once it has confirmed the results of its ongoing overhaul of its product offering and business model, it plans to expand internationally.

The company, formerly known as Focus Telecom, has been backed by Polish private equity fund MCI Capital since 2014. In 2023, MCI became its majority shareholder, buying an 80% stake for PLN 80 million (EUR 18.2 million). It currently carries the stake on its books at PLN 129 million (EUR 29.3 million), valuing the entire company at PLN 161.3 million (EUR 36.6 million). MCI typically allocates PLN 100–200 million (EUR 22.7–45.4 million) to a single transaction, leaving room for further investment.

“With MCI among our shareholders, we are a natural candidate to consolidate the market in the region. We are looking at potential targets, not only in Poland. We are interested in complementary products and customer bases. First, however, we want to confirm that we have created a platform that meets customers’ expectations,” says Błażej Rychlik, CEO of Welyo.

A market leader builds its own platform instead of buying a startup

The Welyo brand was created a year ago through the merger of Focus Telecom and Systell, which had been part of the same group since 2024. The merger brought numerous challenges. Although the two companies offered similar products, they had different platforms, types of customers and approaches to the market. Developing shared technology and an organizational culture while maintaining uninterrupted service took considerable work.

“The benefits more than make up for it. It is not just about cost synergies, but also about joining forces to develop AI tools. Available financial statements show that we generate the highest cloud contact center revenue among providers operating in Poland. We are actively migrating customers to the new platform. That will take several more quarters,” Rychlik says.

By the time he became Welyo’s CEO in the summer of 2025, the global contact center market was already undergoing rapid change. The biggest players were building or buying agentic AI platforms: NiCE acquired Cognigy for almost USD 1 billion. Providers of customer relationship management (CRM) systems, including Salesforce, were expanding into customer service tools.

At the same time, numerous startups emerged that focused exclusively on fully autonomous contact center platforms. Welyo therefore began by looking for ready-made technology it could acquire.

“None of the options met our expectations, however, and after integrating with Systell, we knew how demanding that process could be. So we drew on the expertise of excellent Polish developers and built our own platform tailored to our customers’ needs, using the best available AI models. In 2025 alone, we invested around PLN 20 million [EUR 4.5 million] in it. In the long term, this is a better solution, including when it comes to further development,” Rychlik says.

A global shift in AI pricing

In mid-2025, the company introduced a new offering for the combined group, while retaining its existing business model. It began implementing more significant changes at the start of this year. It introduced a range of AI-powered tools and an AI assistant for customer service representatives. For the past few months, it has also offered customers an AI agent platform.

“In November, we are officially launching the new version of the platform and our offering. We will present it on November 26 at Welyo CX Talks. We have developed the technology to include AI agents that handle repetitive matters independently and pass more difficult cases on to people. This changes both the way work is done and how we price our offering,” Rychlik explains.

Alongside its traditional subscription-based software-as-a-service (SaaS) packages, priced per representative, the company is introducing usage-based billing through “AI credits.” Customers can use them, for example, to pay for minutes of conversations between voice agents and customers or for a number of messages.

“This is a global trend in AI services pricing. We have already signed the first contracts. Over time, we want it to become our main sales model,” Rychlik says.

Expert's perspective

Building a partner network makes sense, but requires preparation

Customers want a fixed, predictable price for a service or product. Once the initial excitement fades, many organizations will begin to feel the problem of uncontrolled costs associated with using LLMs [large language models — ed.].

Passing the variable costs of infrastructure and LLMs on to the customer is tempting, but risky. A simple analogy is the way mobile phone plans have changed in recent years. There will always be someone in the market offering a better, more transparent plan.

Turning to a network of implementation partners in order to focus on the product, on the other hand, is not a new trend. There have been many examples over the years of companies successfully establishing their position with this model. In the enterprise resource planning (ERP) market, for example, Comarch became a top-three player alongside SAP and Oracle by successfully developing a partner network.

The AI revolution undoubtedly creates new opportunities. Companies that can use them to deliver better products faster will have a stronger proposition for partners. This can work, but companies must remember to share their margin so that partners can make money. They also need sufficient resources to give up short-term implementation revenue and make it through the “valley of death”: the period before subscription revenue from partner-led implementations starts coming in.

Using autonomous AI agents effectively saves significant time and money. This is the right direction for software development and, with a few exceptions, the only one. Nature abhors a vacuum. Products will be better, delivered faster and cost less. Sooner than we think, they will displace solutions built using the “old” approach.

At 3HT, projects that would have required teams of several hundred people and at least a year of work just two or three years ago can now be completed by a handful of people in a few months. Understanding the business need and the objective is becoming crucial. So, of course, is the ability to use LLMs and agentic programming while managing risks related to security, personal data and other areas.

Keeping people at the center of customer service

Companies have told Welyo they are concerned about forecasting how many credits they will use. The company says it can estimate their actual needs with reasonable accuracy based on the volume of their customer interactions and historical data from comparable businesses.

“On that basis, we set a minimum amount and the terms for higher-than-usual usage, such as during sales peaks. Many companies in the early stages of transformation want to use AI for everything. But the results do not always justify the costs,” Rychlik explains.

Welyo is pursuing a two-pronged approach. It offers customers an AI agent platform that automates repetitive, routine matters. At the same time, it provides AI-powered tools that make both representatives and the managers supervising them more effective.

“The former can easily generate transcripts and summaries of conversations, for example, while the latter can review potentially every customer interaction from start to finish. We continue to put people at the center, emphasizing collaboration with technology rather than replacing them with it,” Rychlik says.

Good to know

Untapped potential

An AI agent can respond to customers without taking breaks. Welyo has estimated how much potential sales revenue a company could lose from 100 unanswered calls: approximately PLN 1,900 (EUR 431) in the automotive sector and PLN 7,900 (EUR 1,793) in tourism. There are many possible reasons for missed calls, including the time of day — when customers get in touch outside representatives’ working hours — and an inability to expand the team.

XYZ

A new strategy and business model

The new business model is significantly changing the composition of Welyo’s revenue. Both revenue from AI services and the number of customers paying for them have more than doubled since the start of the year. Even allowing for the low starting point, the company sees this as evidence that the market has responded positively to its offering.

“At the same time, we moved away from one-off platform implementation work, which accounted for a significant share of revenue. We now carry it out with partners. We share the margin, but we can focus entirely on developing the technology, which is already producing tangible results. This allows us to offer a better and more comprehensive product, and in turn to increase revenue from our existing customer base,” Rychlik says.

A year ago, Welyo had more than 800 customers and wanted to exceed 1,000 over the following few years. It still serves more than 800, although some customers left as a result of the integration, as the company had anticipated. Welyo believes it can meet its financial targets with a smaller customer base than previously expected.

“We are deliberately moving toward the enterprise segment. We see greater awareness among large companies of what customer service automation can achieve. Previously, we targeted companies with at least a few customer service representatives. Today, the main criterion for us is the number of customer interactions handled, and our platform handles more than 10 million of them a month,” Rychlik says.

Revenue falls during transformation, but growth is expected to accelerate

The market has noticed and welcomed the changes at the company. In his latest report on MCI’s portfolio, Krzysztof Radojewski, deputy director of analysis and advisory at Noble Securities, says Welyo is successfully transforming itself into an agentic AI platform for contact centers.

MCI plans to sell the company in 2029. Since the acquisition, its revenue has grown by an average of 19% a year, largely due to the acquisition of Systell. In 2025, revenue fell from PLN 45 million (EUR 10.2 million) to PLN 42 million (EUR 9.5 million), but normalized EBITDA rose from PLN 15.5 million (EUR 3.5 million) to PLN 16 million (EUR 3.6 million).

“The decline of several million zlotys is a consequence of the transformation under way, but we are already seeing positive results from the changes. According to Fortune Business Insights, the global cloud contact center market is growing by around 17% a year, and the AI segment is growing even faster. We want to return to growth, with artificial intelligence as its main engine. AI offerings still account for a small share of our sales today, but that share will grow rapidly,” Rychlik says.

Global competition and the “SaaS apocalypse” are no longer the main concern

What could hold the Polish company back? Its CEO points out that global contact center players such as Genesys and NiCE have operated in Poland for years. There is always a chance they could suddenly invest more locally, but he does not see this as a key risk. Nor does he consider the emergence of new startups a key risk.

“A year ago, when people were predicting a ‘SaaS apocalypse’ [the replacement of cloud-based software services with in-house solutions built using AI — ed.], I was very concerned about it. But customers value experience and a comprehensive approach to solving their problems,” Rychlik says.

The biggest long-term change for the industry as a whole could be the widespread adoption of agent-to-agent communication — conversations with AI agents on both sides.

“I believe, however, that we are still a long way from that. Humans, supported by AI tools, will continue to play a key role in customer service for years to come,” he says.

Fireside chat

The future of contact centers is not a choice between humans and bots

XYZ: Why does a company operating in dozens of countries need artificial intelligence in customer service?

Dawid Telepski, head of contact centers at LPP: For us, AI’s greatest value in contact centers is the ability to combine scale, speed and quality of service with operational flexibility. It is no longer just about a chatbot answering the simplest questions. We need solutions that can identify a customer’s intent, communicate in different languages, help representatives choose responses and categorize cases correctly. At the same time, they need to recognize when a person should take over the conversation.

At LPP, we use AI both in direct contact with customers and to support our representatives. We can handle several hundred conversations simultaneously, translate communications in real time and automatically route cases to the right teams. Automation is particularly important during sales peaks. We can increase our service capacity as traffic grows, without having to build the entire organization around short-lived peaks.

Are companies that automate customer interactions already gaining a real advantage?

Artificial intelligence may increasingly determine competitive advantage and, in extreme cases, whether a company gives way to more agile players. Ignoring automation can mean not only falling further behind competitors technologically, but above all becoming less flexible, scaling processes more slowly and finding it increasingly difficult to serve a growing business efficiently.

At the same time, I do not see artificial intelligence as a straightforward matter of technology replacing people. The way a representative works is changing, but it is not simply a division between easy cases for AI and difficult ones for humans. AI agents can already handle complex cases from beginning to end. A person steps in when their judgment, a decision or an individual approach is needed.

How are the roles of people and bots changing?

The representative’s role will go beyond taking over conversations from AI. Increasingly, they will work with AI agents as their supervisors: monitoring the quality of responses, training systems, and helping design and develop service processes. They will also make decisions in situations where the AI is not sufficiently confident.

Some customers’ current reluctance to be served by bots may stem from experiences with earlier generations of chatbots, which had limited capabilities and often could not resolve problems effectively. But AI agents are getting better at understanding context, conducting natural conversations and independently bringing cases to a conclusion. I believe that as these solutions develop, talking to AI will become virtually indistinguishable from talking to a person, and customers will become increasingly willing to use this form of service.

Will AI agents replace customer service representatives?

I assume that, in the longer term, AI agents could handle as much as 100% of initial customer contacts. Representatives would primarily receive cases requiring an individual approach or a human decision. For customers, this should not mean a decline in the quality of service. On the contrary, thanks to AI’s speed and availability, they could receive help faster and more effectively.

That is why, instead of talking about the disappearance of the customer service profession, I would speak of its evolution. The future of contact centers does not have to come down to a choice between a person and a bot. I see it as a collaboration between people and AI agents, in which technology expands service capabilities while people contribute their experience, oversee how it works and help create ever better solutions for customers.

Key Takeaways

  1. European ambitions. Welyo, part of MCI Capital’s portfolio, is Poland’s largest provider of cloud contact center technology. Before the fund sells the company, Welyo plans to strengthen its position in the region and is already assessing the market for potential acquisitions. First, however, it needs to complete its ongoing transformation.
  2. AI in customer service. The brand was created a year ago through the merger of Focus Telecom and Systell. In addition to integrating their teams, the companies moved to a new shared platform and expanded their offering with AI tools. Instead of buying an off-the-shelf product, Welyo is building its own from scratch, investing around PLN 20 million (EUR 4.5 million) in 2025 alone.
  3. Results expected soon. The shift in business model led to a fall in revenue from PLN 45 million (EUR 10.2 million) to PLN 42 million (EUR 9.5 million) in 2025. The company nevertheless wants a swift return to rapid growth, driven by AI capabilities, a shift toward enterprise customers and the use of implementation partners.