Mayumi Nakamura Birt, the chief transformation officer of Veltra, spent more than two decades running a software-as-a-service (SaaS) company in the United States. She learned to fail fast, move on and treat mistakes as data.
Six months ago, she came back to Japan to lead transformation at Veltra, the Tokyo-based tours and experiences platform. The weak yen helped. But what really drew her back was something harder to quantify: the sense that Japan might finally be ready to change and a personal hunger to be in the room when it does.
“I like to be part of the game,” she said simply.
Her conversation at WiT Japan, framed around the theme of bridging two worlds, provided a mapping of the gap between how technology companies are built in the U.S. and Japan, and what it actually takes to close it. The bridges she is building are not only organizational. They are cultural, psychological and personal.
The precision trap
When asked if Japan’s culture of precision—the quality that produces spotless streets, flawless service and products that work exactly as intended—was an advantage or disadvantage in an artificial intelligence (AI) world, she said it could become a liability in software development.
“They don’t have a tolerance for failures, often trying to make it so perfect that by the time they get there, the product is already in the past.”
The U.S. model she absorbed over two decades runs on a different logic: ship fast, learn faster, iterate without shame. Neither model is complete. But in a moment when AI is compressing development cycles from months to days, the Japanese instinct to perfect before publishing creates a structural disadvantage. “If they stop to try to make it perfect, somebody else will come in and play.”
Her working compromise at Veltra is 80%. “You have to allow 80%, and that’s OK too.”
Training agents as a Trojan horse for tolerance
Asked for a specific project she’s working on at Veltra in her role as chief transformation officer, Nakamura Birt described how she is using AI implementation not just as a productivity tool but as a change management mechanism, a way to build failure tolerance into the organization without ever framing it as such.
The method: Treat AI agents as new employees and have staff train them. When a human trains another human and something goes wrong, there is shame, loss of face, the weight of hierarchy. When a human trains an agent and the agent fails, none of that applies. “The agent’s not going to feel the pain. They’re not going to cry.”
The longer-term vision is a hybrid workforce—humans and agents operating alongside each other, with the human role shifting toward managing, directing and improving the agent rather than performing the task directly. It is, she noted, a 24-hour workforce.
The parallel to fear of technological displacement is not lost on her. When the internet arrived, travel agents feared for their jobs. Some lost them. Others found that the role evolved rather than disappeared. The pattern, she argued, is consistent: “The change is probably the only thing in this world that is permanent.”
What varies, she said, is whether people choose to be carried by it or crushed by it.
The Hawaii lesson and Japan’s distribution problem
Before joining Veltra, Nakamura Birt worked on a project building a destination management system for Hawaii’s state park network.
The problem Hawaii faced—too many tourists, concentrated in too few places, with inadequate communication about alternatives—is structurally identical to what Japan is experiencing now.
Her solution in Hawaii combined technology with a reimagination of what the destination actually was. Rather than managing one iconic site, the platform introduced the entire state park system as a distributed offering, implemented timed entry across multiple parks and used the booking interface to actively reshape visitor expectations and behavior.
“We managed the entries and then changed the expectation of travelers.”
The result was not just crowd management. It was a reframing of what Hawaii was, moving from a single-destination, single-experience narrative to a layered, distributed one. Japan, she believes, has the same opportunity, but it requires something the Hawaii project also needed first: political will and a shared vision at government level.
“It would be nice to see the government of Japan take a more holistic approach to where they want to see tourists. Right now, you need that heavy lifting from society to say: we need to solve this problem.”
The pain, she suggested, may be what finally generates that will. Nakamura Birt offered a deliberately provocative historical parallel: Japan’s modernization after the arrival of Commodore Perry’s black ships in the 1850s. The shock of external pressure forced open a society that had been closed.
“Japan sees AI as a threat and also sees the labor shortage as a threat. Well, Japan can be very creative to show the rest of the world how it can manage all these challenges.”
The threat, properly harnessed, becomes the engine.
Storytelling as the core competency
Asked what travel companies will look like in 10 years, Nakamura Birt declined the obvious answers—tech company, media company—and chose “storytelling company.”
“AI can tell stories, but humans seek experiences to travel. That story needs to be told. The companies who do well telling that story will utilize AI and other platforms to deliver the right message to the right people in the right way.”
It is a framing that cuts through the noise of the current AI debate in travel. The question is not whether AI replaces the travel company. It is whether the travel company can use AI to become a better storyteller—more personal, more contextual, more attuned to what a specific traveler actually wants to feel, not just where they want to go.
Veltra’s focus on experiences rather than logistics positions it, in her view, on the right side of that shift.
“Experience to learn and see becomes more high-demand in the era of AI, because people will be thirsty for those experiences.”
The women question and the articulation advantage
Two observations about women in the AI era stood out in the conversation.
The first was structural: AI removes the programming barrier. Women who did not study computer science, who were historically excluded from the technical track, now have direct access to building things with technology.
“AI gives power to women who didn’t study programming. You don’t have to be programmers anymore to create something new. It’s a great empowerment.”
The second was cultural and subtler. Women, she argued, tend to be more comfortable listening, asking questions and sitting with ambiguity, precisely the skills that effective AI use demands.
“AI forces us to ask questions. It’s a practice.”
In a culture where asking questions can feel like an admission of ignorance, that reframe is not trivial, she asserted.
Her rapid-fire answer on the most important skill Japan should build with AI was a single word: articulation. Not just asking questions, but being able to describe clearly—to help others put what they imagine into words, to clarify, to specify.
It is, she noted, a skill she had to build herself as a bilingual professional navigating multiple cultures.
“You always want to make sure you’re asking questions and getting the answers correctly. It’s just a habit.”
Japanese women, she said, are ready to build it.
This story originally appeared on WiT.