A strategy can travel across a company in an afternoon. It leaves the boardroom as a tidy sequence of priorities, slides and deadlines. By the time it reaches the people expected to carry it out, it has become something else: a new tool that interrupts a familiar workflow, a manager without a convincing answer, a private doubt about whether this initiative will last. Dima Syrotkin has built his career around that distance. His company, Pandatron, sits between the official plan and the unofficial reality, asking employees questions while change is still happening and showing leaders the patterns that form.
The product is a conversational AI coach. An employee can use it inside the systems where work already happens, talk through a challenge and decide on a next step. The company receives anonymized themes across those conversations. The design makes a wager that corporate change is neither a broadcast nor a software installation. It is a collection of small decisions, repeated by enough people to alter the way an institution behaves.
Syrotkin calls change management “strategy execution.” That plain definition removes some ceremony from a field known for frameworks. It also explains why his work keeps returning to listening. A plan can tell people what leadership wants. It cannot, on its own, reveal where language is confusing, incentives collide or confidence disappears.
“The gap exists in the people, not the plan.”Pandatron’s operating premise
A mission in search of a product
Syrotkin’s founder story does not begin with a cinematic flash of insight. He has said that entrepreneurship started with a broad decision to do something useful, followed by his interests rather than a perfectly formed customer problem. Human development led him toward organizational development. Organizational development led him toward the stubborn problem of change. The sequence matters because Pandatron had to be discovered through practice.
He co-founded the business in 2016, when it was known as Panda Training. The current company came through two major pivots. Asked years later about his biggest accomplishment as an owner, Syrotkin did not name a client, product release or funding round. He named survival: not giving up through those changes of direction. It is a compact answer, but it reveals the texture of his management style. Persistence, for him, is less a slogan than the willingness to let the product change while keeping hold of the underlying question.
Long before the enterprise platform, there was a student organization in Turku. From 2013 to 2014, Syrotkin founded a local AIESEC office and led it from zero to 18 members. His public account notes that nobody dropped out during his term. He spent the following years facilitating leadership programs across Europe. The setting was different from a corporate transformation, but the problem was familiar: a group needs purpose, structure and enough individual ownership to keep moving.
His education added two more lenses. At Turku University of Applied Sciences, he completed a bachelor’s degree in engineering and technology. At Aalto University, he moved into management and international business. The combination helps explain Pandatron’s shape. It treats software as infrastructure, while treating adoption as a question of behavior, trust and organizational design.
The conference clue
The subject of Syrotkin’s 2020 master’s thesis began as a puzzle inside his own startup. Roughly 90 percent of its customers had been acquired through technology conferences and events. Why did brief, in-person contact work better than cold outreach for a young founder selling into an unfamiliar industry? His answer focused on “swift trust,” the rapid confidence that can form when people meet under the compressed conditions of an event.
It is an unexpectedly useful prehistory for an AI-coaching company. Enterprise software is purchased through committees, but adopted through countless judgments about credibility. Employees ask whether a new system understands their work, whether their candor is safe and whether leadership will act on what it hears. Pandatron’s challenge is to produce a version of swift trust without a founder standing across from every user.
The private conversation at enterprise scale
Early on, the Pandatron team expected people to be guarded with a machine. Syrotkin later said the result surprised them: users often seemed more comfortable than expected because they did not feel judged by the AI coach. The finding sharpened the product. A coach that lowers the social cost of admitting confusion can surface the part of a transformation that rarely reaches a status report.
The second half of the system is aggregation. Individual exchanges stay private while recurring barriers become visible to leaders. That distinction is the business model’s ethical and practical hinge. Without privacy, employees have little reason to be frank. Without aggregation, the conversations remain isolated moments of support. Together, they form an organizational listening layer.
Customers named publicly across Pandatron’s history include SAP, Universal Pictures, Skanska, Panasonic, Merck and Mitsubishi. In one account, Skanska said it achieved results comparable with earlier training programs in one-third of the time. Pandatron has also pursued partnerships with established consultancies rather than positioning every engagement as a contest against them. For an enterprise startup, that is both distribution strategy and a concession to how trust moves through large institutions.
The product’s unit of change is deliberately small: one candid exchange, one clearer obstacle, one next action. Scale arrives by repeating that unit without exposing the person behind it.
Research running beside the company
While building Pandatron, Syrotkin has pursued doctoral work at Aalto on structures and processes that decentralize authority and promote self-management. He has cited systems such as Teal organizations and Holacracy as examples of the territory. The research and the startup circle the same tension. How does leadership set a direction without reducing everyone else to passive recipients?
His writing supplies another view of the operator. On Medium and LinkedIn, Syrotkin has published about B2B sales, fundraising, working less as a CEO, startup failure and books. His advice to founders is blunt enough to fit on an index card:
- Sell before you build.
- Build before you hire.
- Hire before you raise.
The order came from experience rather than doctrine. When Pandatron closed a $621,000 pre-seed round in 2021, Syrotkin wrote that the company had raised $882,000 in total. He also warned that investment was a dangerous proxy for success. Capital could accelerate a working engine or deepen the fall of a company without product-market fit. It is the kind of line a founder writes after learning that announcements are easier to produce than durable demand.
There is a similar correction in an essay about working less. Syrotkin described moving from an overworked chief executive toward a strategist leading a 12-person team. Delegation, later interviews suggest, became central to recruiting capable colleagues and getting out of their way. The shift echoes his academic interest in distributed authority: the founder studying self-management also had to stop making himself the center of every decision.
“Artificial Intelligence should be used to foster human development.”Dima Syrotkin, 2025
AI that ends the conversation
Syrotkin’s preferred future for AI is not a machine that maximizes time spent with it. He has argued that it should help a person find the next thing required to unlock more of their potential. That creates an unusual product standard. A useful coaching session should end with the user doing something in the world, not remaining inside the chat.
This principle is becoming more consequential as Pandatron shifts from the language of learning and development toward “change activation and execution intelligence.” The company now presents itself as infrastructure for making strategy real. Its team includes change operators and software builders, while its public group of framework partners includes scholars of psychological safety, behavior, choice and power. The message is that a language model alone is not the method. The method is the structure wrapped around the conversation.
In 2025 and 2026, Syrotkin’s interviews increasingly centered on AI adoption itself. The irony is useful: companies are trying to deploy AI, and Pandatron uses AI to address the organizational resistance that deployment creates. The tool becomes both the subject of change and part of the change process.
The institution that can notice itself
Syrotkin has described his broad ambition in terms of self-awareness. The newer version is institutional: he wants companies to reinvent themselves repeatedly instead of holding a successful shape until it becomes a trap. Those aims meet in Pandatron’s basic mechanism. A person reflects. Many reflections reveal a pattern. The institution adjusts.
His own path has followed the same loop. A student chapter became an early laboratory for leadership. Conference sales became a thesis about trust. Two pivots became a lesson in persistence. Doctoral questions about authority traveled into a product that gives employees room to think. The clean narrative only appears in retrospect; the lived sequence was a series of clues.
The unresolved question is the interesting one. Can an AI-mediated conversation preserve enough trust to help a person speak honestly, while producing enough structure to guide an enterprise? Pandatron’s work with large companies suggests the mechanism can operate across languages, industries and geographies. The longer test is whether leaders use what they hear.
A strategy deck is a declaration. Execution is a feedback loop. Syrotkin’s project is to shorten that loop until a company can notice hesitation while there is still time to respond. The software speaks, but the product is attention.