MGS 2026 Recap: MyRealTrip's AI-Native Organization

You've rolled out AI tools across the team, so why does the way everyone works still look exactly the same? If work still moves from planning to design to engineering to QA in the same order, the time AI saves just disappears into waiting for the next handoff.

The daglo marketing team attended Modern Growth Stack (MGS) 2026, hosted by AB180 at the Westin Seoul Parnas from July 21 to 24, 2026, under the theme "Stack AI, Rewrite Everything." This recap covers "Rewrite Travel: Redesigning a Travel Platform with AI," presented by Donggun Lee of MyRealTrip, one of Korea's largest travel platforms. We transcribed it with daglo and pulled out what practitioners can use.

An AI-native organization doesn't just bolt AI onto existing work. It redesigns roles, performance reviews, and customer touchpoints on the assumption that the cost of building is approaching zero. MyRealTrip merged all of its product roles into a single Product Engineer role, and AI now handles the first response to 80% of its customer chats.

Where an AI-native organization starts: one Product Engineer role

MyRealTrip's shift to an AI-native organization began with merging job roles. Customer problems don't arrive split by job function, so the idea is to let the person who spots a problem solve it from start to finish.

In the mobile era, a problem passed through roughly eight roles, not because that was ideal but because nobody could be an expert in everything. As AI lowered that barrier, MyRealTrip merged roles in three stages.

The CEO went first. Acting as a Product Engineer, he built a restaurant discovery service and shipped it in the real product. After three months with little response, he took it down. Because it had been easy to build, it was easy to pull. When building gets cheaper, so does walking away.

From job functions to customer problems: two examples

Once roles merge, work no longer starts with "What's my frontend task today?" It starts with "A customer says this is frustrating." The session shared two flight-related examples where one person carried a customer problem all the way through.

• Why keep searching for flights? Travelers without fixed dates or destinations browse with a vague "anywhere good in July?" until a deal like Hong Kong for 180,000 won catches their eye. After seeing a competitor offer this, MyRealTrip's head of marketing built the feature end to end and shipped it. It is still an important feature today, even under a new owner.
• Flight change alerts: Asked why they used an airline's own app even though it cost more, customers said the airline app keeps sending push alerts when a flight changes. A backend engineer on the flights team switched to the Product Engineer role and built the feature alone, from design to launch.

Why MyRealTrip closed its AI Lab and opened a Sales Lab

MyRealTrip shut down the AI Lab it created in 2024 two months before the session. Leadership decided everyone's AI skills had reached a level where AI literacy no longer needed its own team.

The next question was: if building moves into AI's territory, what has to stay human? The answer was meeting customers on the front line. The company calls it sales for convenience, but it means every kind of direct customer contact.

MyRealTrip started sales training for Product Engineers, and an engineer from the accommodations team moved into accommodation sales, meeting hoteliers and owners of small motels and guesthouses in person. When an owner says the admin tool eats 30 minutes to an hour of a busy day, the answer is, "Then let me fix it for you right here." Owners have called it "like magic," and one said it was the first time they had ever met a developer. Nine engineers have now moved into sales through the Sales Lab.

AI in customer service: 80% first response and a top satisfaction ranking

At MyRealTrip, AI gives the first response to 80% of chat inquiries and closes 50% without a human. On the phone, AI handles 57% of calls from start to finish.

Travel is a high-stakes purchase, often millions of won, so cancellations, refunds, and changes matter enormously. The company receives about 75,000 inquiries a month. After giving AI direct access to internal data so it could process cancellations and refunds on its own, the numbers looked like this.

Human agents now take only the truly difficult cases, so each agent handles more and agent satisfaction has gone up too.

Questions to ask before building an AI-native organization

MyRealTrip insists that every change prove itself in revenue and operating profit. Revenue per employee is a core metric, and "AI Nativeness" in individual reviews is judged by the revenue and profit someone created, not by how much they use AI.

The session's conclusion was that what matters now is how much a small team, or even one person, can own and deliver. To apply this to your own team, start with these questions.

• Handoffs: Which handoffs exist only because someone lacks a specific skill?
• Exit criteria: Have you decided in advance when to shut down a feature or experiment that isn't getting traction?
• An end date for AI teams: Does your AI team or training program have a clear point at which its job is done?
• Reviews: Do you evaluate AI by the results it produces rather than by usage?
• Org metrics: Are you tracking output per person instead of headcount growth?

For recordings where it matters who said what, like multi-speaker sessions or Q&A, daglo's Speaker ID is a useful feature. It remembers the voices of the people in your recordings and fills in their real names instead of "Speaker 1" from the next recording on, and daglo is the first service in Korea to offer it to every user. Learn more about Speaker ID

For the other sessions we attended, see our full MGS 2026 recap. Event details are on the official MGS 2026 site.

#MGS2026 #ModernGrowthStack #MyRealTrip #AINative #ProductEngineer #AICustomerService #GrowthMarketing #daglo

👉 Transcribe your next session