Based on a conversation with Dima Syrotkin, founder of Pandatron.

What is change management? Effectively, strategy execution. And right now Pandatron is focused on the most pertinent strategy of our time: AI. 95% of AI pilots fail per MIT report. That costs 28B+ USD.

Today, his company Pandatron is writing the new playbook on change management. Its AI-coach brings personalized support to each employee, while anonymized reports give leaders an accurate sense of what’s actually happening on the ground during organizational transformations.

The Innovation: Managing Change Better With AI Coaching

70 to 80% of corporate change management programs fail. That’s a stunning percentage; the combined yearly spend on these programs numbers in the billions of euros. The cause is not technological. Consider the case of buying and implementing a new CRM system, or - today's hot topic - realizing AI adoption throughout an organisation. No matter how good the software or the technical aptitude of the implementation consultants, most of these projects fail due to pushback from employees. 

To tackle this problem, which Dima refers to as “ignoring the human side of innovation”, Pandatron has built an AI change agent. Rather than broad datasets, it’s trained on exclusive content from top experts, therapists and coaches, and further tuned to the specifics of each company.

The key functionality is two-fold:

1. The AI model speaks with each of the thousands of employees that make up the enterprises of this world during periods of change. It helps them understand why a transformation is taking place, alleviates concerns, and provides individualized coaching along the way.
   
2. It gathers data from these conversations, anonymizes and analyses it, and provides reports to the management layer. This helps leadership better understand where people are struggling, where they need more support, and what's at the core of the challenges.

"We already know that the old playbook - endless email chains, impersonal trainings, pretty powerpoint presentations - doesn’t change behavior, so many enterprises stopped investing altogether. But that’s the wrong lesson."

"AI introduces lightweight tools that deliver outsized impact for the cost. Without it, some of the boldest innovations will remain a slide in a deck instead of becoming reality. I can’t let that happen."

Real-World Impact: Democratizing Access To Professional Coaching   

Pandatron is already seeing major usage in the real world across a swathe of industries. Their customers are large enterprises that deploy their change management solution to support hundreds of thousands of employees.

Skanska, one of the largest construction companies out of Sweden with global operations, told Dima that they achieved similar results to previous corporate training programs, only three times faster. Other notable names include Panasonic, Merck, SAP,  Universal, and Mitsubishi.

Practically unlimited scale is another major advantage for Pandatron. “The accessibility of our technology, in terms of cost and usability, is a game changer (pun intended) for the change management industry”, says Dima.

“Historically, coaching and executive were typically used in the same sentence due to the premium prices professional coaching services fetch. Thanks to AI, we‘re suddenly able to provide the same level of support to thousands of people across organizations - without significantly raising the total cost.”

Positioning this offering as a challenger to classic management consultancies like McKinsey would likely raise eyebrows and shut doors. So Pandatron took a different path: they opted for a growth strategy that relies on partnering with these globally entrenched firms instead. It’s proven successful, as their solution is already in use across North America, Western Europe, and even Japan.

Better Development Through Listening

In 'The Listening Society', author Hanzi Freinacht describes how societies and governments need to pay attention to how the psychological development of their citizens is evolving. In very simple terms, it advocates for tools like therapy to be in every school and organization. This idea really shaped Dima’s thinking. “I feel that it’s just so important, and in many ways what we're doing with our product is kind of very similar,” Dima muses. “We’re just applying it to a very particular problem.” 

Perhaps the most crucial realization for Pandatron was counterintuitive. “We expected people to be hesitant to speak on their challenges with a machine.” Dima recalls, “Instead we found that they actually feel way more comfortable than when speaking to a human. People don’t feel judged by Panda [the AI coach], and so they have an easier time opening up and discussing their true concerns and challenges. This was frankly mind-blowing, and made us realize that we were sitting on some real gold."

Another concept that increasingly occupies his mindspace is ethical behaviour, even in for-profit companies. And it’s especially important in the AI space. For example, Dima swears by Claude. Not just because he thinks it’s a better model than ChatGPT, but also because Anthropic appears to be a highly principled company.

“Artificial Intelligence should be used to foster human development”, Dima explains, “It shouldn’t be engineered to maximize usage, but rather help you find the next thing required to unlock [more of] your potential.” That belief is foundational to Pandatron’s product. It focuses on coaching through asking people questions and helping them reflect.

Creating Resilient Institutions

Dima sees the tremendous government support as a major advantage for European startups, one that American counterparts don’t have. Startups more often than not fail, so tapping into the resources offered by the EIT AI Founders Club was a prudent and logical move. Both the community and funding have made a noticeable, positive difference for both his personal and his company’s growth. 

Looking ahead, Pandatron’s founder and CEO envisages a feature where companies don’t spend decades in the Fortune 500 only to be suddenly outcompeted by newer, more innovative companies. Instead, he wants to help create institutions that really not just survive but also thrive, that can successfully reinvent themselves time and time again. Something that he hears is deeply needed, especially in Europe.

Learn more about Pandatron: Website | LinkedIn 

When Diana Miron noticed how many kids lost interest in coding before they ever wrote a line of real code, she decided to change that. Together with her co-founder, she built Bit&R: an AI-driven platform that teaches kids between the ages 10-14 to code by creating and sharing their own games.

With their soon to launch AI co-pilot 'Robert', the team is making coding playful, inclusive, and accessible for young learners everywhere, paving the way for a future where digital literacy is as natural as reading or writing. In this story, we speak with Diana about their vision and innovation, the impact they are making and the choice to join the EIT AI Founders Club.

The Innovation: Personalized learning paths

Tweens (kids aged 10–14) often sit between block-based tools and adult coding platforms, with few options that match their skill level and pace. Most existing tools fall at either extreme: overly simple drag-and-drop environments that lack substance, or complex coding environments that assume prior knowledge and maturity.

When results take weeks, attention fades. So the goal for Diana and her co-founder Claudia was enabling children to build something real and fun within hours. That mission crystallized into Bit&R: an adaptive, AI-driven platform that lets kids learn by creating and sharing games.

At its core, Bit&R offers interactive game-based learning coding projects, a game building feature and a collaborative community. “We are now working on developing our own generative AI model, 'Robert', to act as the kids’ co-pilot, while also personalizing the learning journey for each child" says Diana.

Key features that make their platform stand out are:

• 360-degree learning experience - the platform uses real code that produces playable, community-sharable games; far beyond static challenges.
• Proprietary AI-copilot - 'Robert' guides learners in real time and makes coding feel like play rather than chore.
• Fuzzy logic and multi-linguistic semantic analysis - ensures curated, child-safe educational content.
• Translinguistic algorithm - interprets text independent of language and enables seamless scaling across languages and contexts.

"What sets us apart is the speed and playfulness of the experience. A child can build a website or a game in just a few hours, gaining confidence and motivation, while also being part of a community where they can showcase their creations."

Real-World Impact

Bit&R has already served over 1500 users, both as individuals at home and in ICT classes in 17 schools across Europe. Their main users are children aged 10–14 that use Bit&R to learn at home, at their own pace and in their own time while also providing value for homeschooled students. They’ve also reached coding hubs to future use Bit&R as a third-party coding platform integrated into their course curriculum.

Beyond just teaching skills, Bit&R contributes to inclusivity by making coding education accessible to children regardless of their background, preparing them for a future where digital literacy is as fundamental as reading and writing. The company is partnered with local NGOs that use Bit&R to teach coding, extending their mission beyond technology to empower and impact underserved communities.

"A standout example is David, one of our top users on the Bit&R platform. His passion for coding and determination were evident early on, and we took him on as an intern at just 17 years old. Within three months, he grew into a proficient full-stack developer, showcasing the potential of Bit&R and accessible coding education."

From self-taught techie to Edtech founder

“A book that I really enjoyed is Hackers and Painters by Paul Graham,” says Diana. “You can see coding as a creative act: something that anyone can use to build, express, and solve problems. My journey into entrepreneurship came from a desire to make coding more accessible, engaging and fun," she recalls. “I co-founded JSLeague as a B2B tech and coding training business, with Bit&R as our spin-off product for kids. Together with my co-founder Claudia, we launched JSKids in 2021 - an online coding class designed to teach children practical coding skills through the engaging experience of game development. This approach not only kept kids entertained but also redefined how they spent time in front of screens during the pandemic.”

The team saw a significant gap in coding education for tweens. These young learners wanted to move beyond block-based coding challenges, but struggled to find fun, structured resources that matched their skill level. This motivated Diana and Claudia to create a solution that could bridge this gap.

"Over time, I’ve learned that it’s less about chasing a perfect idea from the start and more about staying adaptive, listening carefully to users, and creating products that help children prepare better for the future."

They intentionally designed Bit&R as a scalable tech product: accessible to every kid on Earth, with empowerment as the focal point. Through this approach, it fosters creativity, diversity, and innovation in tech education.

One of the most common misconceptions about Bit&R is that it’s just another coding platform. In reality, their AI co-pilot could create an entirely new way of learning that feels natural and engaging. Inside the team, the culture is collaborative, curious, and deeply mission-driven. They are constantly learning and exploring new things in both tech and business. “We combine deep technical expertise with a human-first approach to education, ensuring that as we scale, we don’t lose sight of our purpose.”. 

Scaling the platform: code education everywhere

The team invested hugely to accommodate various learning styles and cognitive needs and making the platform a robust, forward-looking solution for preparing all children for a digital future, with a retro child-friendly interface that requires no prior coding experience.

“Our scaling strategy begins with English and German-speaking countries in Europe, where we already have traction, and expands globally through schools, and partnerships with other coding hubs.“ says Diana. 

What makes Bit&R adaptable across geographies and industries is its core AI co-pilot, which can adjust learning paths and scales quickly across languages. This flexibility ensures that the solution can grow and remain relevant across multiple contexts.

Joining the EIT AI Founders Club was a strategic step in Bit&R’s growth journey: a way to learn from peers, exchange expertise, and strengthen their approach to responsible scaling.

"I decided to join the AI Founders Club to connect with like-minded entrepreneurs and accelerate Bit&R’s growth by sharing knowledge, challenges, and AI best practices, especially when it comes to safety and ethical AI development. The program fits perfectly into our vision of scaling responsibly while keeping innovation and impact at the core of everything we do."

Looking Ahead 

The next steps for Bit&R include launching the AI model with Europe as a test-bed and expanding it in the future to support multiple coding languages and STEM concepts. In terms of business growth, the team plans to scale globally, offer support for over 20 languages and teach additional age groups how to code.

“Our long-term ambition is to build the first AI model designed for kids, bringing tech innovation into EdTech, while redefining how kids of all ages learn to code. We believe digital literacy should be as fundamental as reading or writing” Diana affirms.

Learn more about Bit&R: Website | LinkedIn 

A chat with Rens Dommerholt, Bekk Blando, & Kaspar Rothenfusser, founders of Thoughtful Oasis 

Most startups begin with a clear industry itch to scratch. Thoughtful Oasis started with something more abstract and far more ambitious: a shared obsession with search. 

Co-founders Rens Dommerholt, Kaspar Rothenfusser, and Bekk Blando didn’t set out to launch an AI company. In fact, none of them wanted to. What pulled them together was a simple but profound question: how do you find something if you don’t know how to ask for it? 

The Innovation: A New Way of Searching 

At its core, Thoughtful Oasis is building a new kind of search engine. One that doesn’t depend on exact words or neatly phrased queries. Because that is the problem with conventional search: you can’t find what you don’t know how to name. Type in the wrong word, miss the jargon, or search for a ‘vibe’, and existing tools fall short.  

Thoughtful Oasis tackles this with what they call vector ontologies: a retrieval and embedding system that automatically creates a taxonomy from any dataset and uses it to map the qualitative and quantitative features that matter. 

Imagine asking for ‘a playlist for a road trip when you’ve just broken up’ or ‘which evacuation protocol to use if a flood threatens a village built below sea level.’ Instead of returning irrelevant results or empty searches, the system finds the most fitting answers by interpreting intent, context, and even vibes. 

This approach enables: 

• Asymmetric search – finding relevant answers even when the query and result don’t use the same words. 

• Multi-shot search – building on previous queries seamlessly. 

• Qualitative search – tapping into attributes like style, mood, or condition. 

• Auditability – every result comes with a traceable explanation. 

“Almost all problems boil down to finding the right thing at the right time,” says Dommerholt. “That’s the frontier we’re working on.” 

Real-World Impact: Helping Firefighters Think Faster 

The team’s first demos were playful. Kaspar classified music tracks by “energy,” “mood,” and “liveliness,” allowing users to search for songs based on feelings rather than titles. Later, they applied the same method to clothing, where typing “business casual interview at an AI startup” would surface fitting outfits. 

But the technology’s potential came into sharp focus in their pilot with the North and East Gelderland Safety Region in the Netherlands. Here, Thoughtful Oasis developed a decision-support tool for firefighters and incident commanders. 

“During a fire, commanders have on average six minutes to make decisions”, explains Dommerholt. “The assistant integrates land registry data with reports from the control room, then searches through official training materials and protocols using the national incident taxonomy. Within this crucial time-limited response window, the assistant can now deliver actionable advice drawn directly from training manuals.” 

The results are concrete and actionable: 

• If a pre-1977 house with a corrugated roof catches fire, the system instantly flags asbestos protocols. 

• If an industrial building is involved, it retrieves the relevant containment procedures before the commander even asks. 

Because the advice is drawn directly from the resources that trained commanders, every suggestion is verifiable and trustworthy. The system leaves behind an audit trail - a critical safeguard in mission-critical environments where accountability is non-negotiable. 

The pilot has already shifted mindsets. “It’s built trust and understanding within the Safety Region about the responsible use of AI,” says Dommerholt. “And it proves that a black-box RAG model could never deliver the same reliability.” 

Looking further, the stakes are high: within 5–10 years, half of the firefighting workforce in the Netherlands will retire. Tools like this could help maintain service levels with fewer personnel by speeding up decision-making, reducing risks of escalation, and freeing firefighters to handle more incidents. 

Three Paths Converge on Search 

Behind this technology stands a team whose collaboration is as deliberate as the system they’ve built. 

The story of Thoughtful Oasis began inside the Novel-T incubator in the Netherlands, where Rens Dommerholt worked supporting early-stage founders and university spin-offs. For years, he helped entrepreneurs refine their strategies, hire teams, and scale up their businesses.

“Being around entrepreneurs that closely, you eventually realize you are one of them too,” he reflects. “I’m not the first idea guy, I’m the person you call when you want to make it a thing.” 

That’s how he met Kaspar Rothenfusser, who at the time was developing his first startup. Rens coached him through that early venture and helped it reach revenue, but he also recognized that Kaspar’s analytical mind was capable of tackling much more complex challenges. He encouraged him to start something new, something worthy of his technical depth. That conversation became the seed for what would later evolve into Thoughtful Oasis. 

Kaspar spent the following years exploring how large language models process and interpret data, eventually uncovering a potential breakthrough in how machines could search for information. When he presented this concept at an AI builders’ community event, it caught the attention of Bekk Blando, an experienced developer specializing in graph and search technologies. 

Bekk immediately recognized the idea as a solution to technical problems he’d been struggling with. “Can I buy it?” he asked. Kaspar smiled and replied, “You can use it if you help me build it.” 

With Rens joining as the pragmatic builder and strategist, the trio officially came together in early 2025. A team bound by curiosity, ambition, and the shared conviction that search is one of technology’s last great unsolved frontiers.  

Building Beyond Borders 

Though founded in 2025, Thoughtful Oasis has already caught the eye of industries where search is mission-critical: retail, government, and R&D. The company operates on an API model, with design partnerships to refine and stress-test the engine in complex settings. 

Joining the EIT AI Founders Club was a deliberate step to expand horizons. “We already had a strong network in the Netherlands and some access to the U.S.,” says Dommerholt. “The EIT AI Community offers us a chance to see what other European founders are building. Besides that, EIT is a great organization supporting many different startups, not only with funding. Now that we’re part of this inaugural community, we get front line access to new initiatives and support systems. The AI Community makes it very easy to make good use of the existing infrastructure, and being coached by Fabrizio Della Pace is very useful for making sure we’re on the right track.” 

Looking Ahead 

The team is now preparing to wrap up design partnerships and move toward a hyperscalable, generic solution. Select CTOs already have access to the playground API, and more industries are being explored. 

The long-term ambition? Nothing less than solving one of humanity’s most fundamental problems.  “We truly believe search is the biggest unsolved challenge out there,” says Dommerholt. “If we can fix that, the impact will last for decades.” 

Learn more about Thoughtful Oasis: Website | LinkedIn