Karel Callens and the distance to an answer

In 2019, Karel Callens built an Amazon Alexa skill with his analytics platform behind it. He could ask how much committed annual recurring revenue the company had. The assistant returned a short answer, then emailed a fuller report with a dashboard link for further investigation. There was something agreeably domestic about the arrangement: a business question, spoken aloud, followed by the paperwork.

He expected voice interfaces to change analytics sooner than they did. Years later, he was willing to say so publicly. The experiment worked; the anticipated change in habits arrived more slowly. Anyone who has ever watched an office cling to an old spreadsheet will appreciate the distinction. Software can be ready well before the people using it choose to move.

That gap runs through Callens's career. The Leuven-based CEO and co-founder of Luzmo keeps returning to the space between possessing information and being able to use it. His question has become less about which screen will win and more about whether an answer remains dependable as it travels between screens, assistants and applications. The Alexa experiment was an early attempt to shorten that journey.

A consultant encounters the waiting room

Before he was selling a way around complicated analytics projects, Callens worked inside them. His background includes business intelligence and data visualization at CSC and a big-data and analytics role at Belgium's federal economy administration. He studied commercial engineering at KU Leuven and management of innovative organizations at the Catholic University of Louvain. The combination put business questions and technical delivery within the same field of view.

By 2014, Callens and Haroen Vermylen were working together in consultancy, largely serving bigger corporations. Conversations with clients and acquaintances exposed a smaller-business problem: companies wanted useful analysis but could not afford the technical support required by existing systems. Development of what became Cumul.io began in March that year. The company was founded in 2015; the platform launched in April 2016.

In those early years, Callens also handled frontend development and design. The future chief executive was involved in the part customers would actually touch. An interface that demanded another expert, another handoff or another explanation would have recreated the very obstacle he and Vermylen had set out to remove.

English, even when everybody spoke Dutch

Callens's ambition crossed borders early. He explained in 2021 that the company had used English from the beginning even when everyone spoke Dutch. Belgium, the founders believed, was too small to be their whole market. Choosing a working language was one modest, practical way to make that belief operational. International expansion began before the office needed an interpreter.

Customer contact was equally deliberate. The first hire worked in customer success. Callens described a team in which everyone encountered customers, giving product decisions some exposure to the people who would live with them. The first customer departure hurt, he acknowledged, before becoming a lesson about the balance between listening and following a direction.

There was also a stated cultural value: “be personal, globally.” Callens linked it to respecting different contexts and accepting that founders could learn from their colleagues. These are useful admissions from someone in the answers business. A company selling better decisions ought to have room for an employee to point out that the chief executive has reached the wrong one.

Buying the room to compete

Money entered the story as a question of freedom. At a Scale-Up Marathon discussion with Billit's Tom Van Gaever, Callens described outside capital as the means to take the steps needed to scale. Van Gaever approached freedom through independence from investors. Their different answers exposed how personal the financing decision can be, even when both founders are looking at spreadsheets.

Callens said Cumul.io had initially bootstrapped, then needed external investment as large competitors entered its market. He also warned that a funding round could take a year from preparation through negotiations. Urgency, in that account, weakens a founder's bargaining position. Investment buys possibilities, but securing it requires its own supply of time.

On January 17, 2023, Cumul.io announced a €10 million Series A led by Hi Inov-Dentressangle, with Axeleo Capital, LRM and SmartFin participating. The announcement described more than 240 SaaS customers and a team of over 40. More than a third of revenue already came from North America. Those figures belong to that moment, when the company was planning further hiring and product development.

A name that needed less explaining

Eight years after the company's founding, even its name had become a small exercise in friction. Cumul.io invited alternative pronunciations. Callens called it a tongue-twister, acknowledging that a useful conversation starter had become an awkward fit for a business selling speed and clarity. There is a limit to how much explanatory work a dot and a few syllables should require.

In September 2023, Cumul.io became Luzmo. The new name drew on the Spanish word for light, with “mo” carrying ideas of more and motion. Brand consultancy GoodLove worked with the team on the change. Callens publicly credited the people involved rather than presenting the new identity as a solitary flash of inspiration.

The rebrand also forced a discussion about what customers were buying. Faster integration mattered, but so did the experience of understanding something. A reporting module might satisfy a product checklist while leaving its users baffled. Giving the business a clearer name was one expression of the same ambition that had produced it: reduce the work required to get to the point.

“Our previous name, Cumul.io, was a tongue twister.”

Karel Callens · On the 2023 rebrand

The feeling behind the figures

That September, Callens wrote about the emotional side of analytics. His concern was the difference between the promise of insight and the actual experience of using analytical software. Numbers can be correct and still leave someone unsure what to do next. Access alone does not provide confidence; a person needs enough context to interpret what is in front of them.

This helps explain his interest in customer-facing analytics. A software company may analyze its own business carefully while giving customers a much poorer view of theirs. Callens's argument asks product builders to treat the customer's understanding as part of the product experience. The person looking at the chart has a job to finish and little reason to admire the infrastructure.

At SaaSOpen in New York in March 2023, he presented five levels of customer analytics experience. They moved from basic shared information through personalized dashboards, native integration and interactivity to proactive reporting. The framework offered product teams a way to discuss progress beyond adding another chart. Each step gave the customer more scope to investigate, act or receive something useful without asking.

When simplicity needs an escape hatch

Callens has also described the limits of his own company's approach. In a November 2024 article, he explained that drag-and-drop tools could become too restrictive as customers' requirements grew more specific. A simple editor is welcome until the thing you need cannot be expressed inside it. At that point, its convenience begins to feel rather like a locked door.

Luzmo's response was Flex, a software development kit that reused the platform's backend while allowing developers more freedom over the frontend. Callens described the design work as separating necessary flexibility from burdensome complexity. Connections, queries and infrastructure could be handled behind the scenes; developers still needed room to shape the experience their customers would see.

His discussion of AI carried a related concern. People should be able to understand how a visualization was produced and correct it when necessary. A generated chart has to survive inspection. That is a more demanding measure than whether an assistant can produce something plausible on request, and it gives the user an active role after the first answer appears.

Another curve before the plateau

By April 2025, Callens was discussing a different kind of design problem: how a growing company prepares for its next phase. His SaaSiest conversation centered on stacking S-curves, including changes to the ideal customer profile and the organizational consequences of pursuing a new audience. Sales, marketing, product and customer success all have to understand the direction.

An S-curve is a useful picture because it gives growth a shape other than a permanently rising straight line. Expansion can flatten. A new opportunity can require work while the existing business still needs attention. The discussion put Callens's task as an executive alongside his task as a product builder: make the next move without losing the people already depending on the last one.

The question outlives the interface

Luzmo now describes its offering across products, chat and AI agents. Callens's recent public reflections are consistent with that broader direction. He has argued that useful answers can take different forms and that definitions, permissions and reliability must travel with them. His old Alexa skill looks less like a failed prediction than an experiment whose central question remains open.

His candor about timing is part of the appeal. He made a working tool, formed an expectation and later revised it. A founder can learn something from a useful experiment even when the market declines to follow the proposed schedule. Customers have their own calendars, and they rarely consult the forecast before opening their inbox.

The ambition that emerges is practical: people should be able to ask a question where they already work and receive an answer they can examine and use. Callens began with the frustration of analytics projects that asked too much of their users. Years of product work, fundraising and changes of direction have kept bringing him back to that distance. He is still trying to make it shorter.