Icana.AI founder Erik van Eekelen bootstrapped his way to his first 10 clients as a solo founder using the very AI tools he was building. Here is what he learned along the way.
Technology alone isn’t enough. Successful AI products require strong alignment between technology, human workflows and organisational culture
Erik van Eekelen had spent more than twenty years watching organisations try to adopt new technology before he decided to build something himself. The problem he kept seeing was always the same.
Customer-facing teams were having thousands of conversations every day, but almost none of those interactions were being reviewed, analysed or used to help people improve. Traditional quality assurance processes typically examined only a tiny fraction of calls. Insights were lost, feedback came too late and the opportunity to coach people in real time was being missed entirely.
“That gap inspired me to build Icana.AI,” he says. “I wanted to create technology that could analyse 100% of customer interactions and turn them into practical coaching insights that help people perform better.”
The result is CallCoach, Icana.AI’s flagship platform, built on a simple premise. Every customer conversation should be an opportunity to learn, improve and deliver better outcomes.
The gap
Van Eekelen is careful to point out that existing tools in the market were not without value. Transcription and sentiment analysis had their place. But they were only scratching the surface of what was actually happening in a customer conversation. “We look much deeper,” he says. “CallCoach evaluates elements such as tone, pacing, articulation, empathy and conversation structure.” The platform uses a multi-model architecture, combining several large language models with voice analysis and behavioural scoring to interpret interactions more holistically.
The difference in coverage is significant. Where most quality assurance processes review a sample, CallCoach analyses everything. That completeness, van Eekelen argues, is what gives organisations real visibility into how their teams are actually performing rather than a partial picture that may not reflect the full reality.
One feature he is particularly proud of is the CallCoach Simulator, which allows agents to practise realistic customer scenarios, including actual calls they have previously struggled with, in a safe environment. They receive the same AI analysis they would get from a live call, creating what he describes as a powerful feedback loop between training and real-world performance.
Building with focus
Van Eekelen initially bootstrapped the business as a solo founder, using AI systems extensively to support his own operations. That approach allowed him to reach his first ten clients and build a healthy revenue base before bringing in additional resources.
He describes the decision to focus narrowly as one of the most important he made early on. Rather than building AI for its own sake, he concentrated on a very specific, real-world problem: conversation intelligence and coaching for sales and customer service teams.
“That focus allowed us to deliver tangible value quickly,” he says.
The platform was also designed from the beginning to evolve with its users. Rather than treating CallCoach as a fixed product, van Eekelen built it as a scaffolded system continuously calibrated with input from experts, users and multiple large language models. Working closely with customers during development meant the insights remained practical and grounded in how teams actually operate rather than how they are assumed to.
Going deeper than other tools
A core part of van Eekelen’s philosophy is the belief that AI should make work more meaningful rather than simply automate it away. He is direct about what that looks like in practice.
“I’ve always believed that AI should empower people rather than replace them,” he says. “If we design it well, it can remove repetitive work and give people better feedback so they can focus on the human side of their roles.”
That philosophy shapes how Icana.AI approaches innovation. Rather than chasing every new capability as the AI landscape shifts week to week, van Eekelen keeps the focus on the real problem the business is solving for customers. Discipline and focus, he says, are essential in a field that can easily distract founders with what is new rather than what is useful.
Lessons from the hard parts
Building trust in AI systems has been one of the more persistent challenges. AI can be powerful, van Eekelen acknowledges, but it can also behave in ways that are difficult to predict. For organisations to rely on AI-generated insights, they need confidence that the system is fair, consistent and calibrated to their specific environment.
To address that, Icana.AI invested heavily in transparency and calibration, actively involving clients in refining how scoring and feedback work within their own contexts. That collaborative approach has been central to building the kind of trust that turns early clients into long-term partners.
The broader lesson van Eekelen took from those experiences is one he applies across the business. “Technology alone isn’t enough,” he says. “Successful AI products require strong alignment between technology, human workflows and organisational culture.”
The key is to remain curious, keep learning and maintain a clear sense of purpose about why you started in the first place
For entrepreneurs at the beginning of their journey, van Eekelen’s starting point is straightforward. Focus on solving a genuine problem that people actually care about. Technology trends come and go, but businesses succeed when they address real challenges in meaningful ways.
His second piece of advice is to start small and move quickly. Modern AI tools have dramatically reduced the time and cost required to turn an idea into something usable, meaning a small team can prototype, test and refine products far faster than was possible even a few years ago.
And if things start to slow down, he says, it is worth asking two questions. Are the most valuable things actually getting done first? And can internal processes be accelerated with AI agents?
“The key is to remain curious, keep learning and maintain a clear sense of purpose about why you started in the first place,” van Eekelen says.
Keep up to date with our stories on LinkedIn, Twitter, Facebook and Instagram.
