Greetings from the Mercan Editing Team.

Over two days in February, approximately 60 members of the AI Task Force* (“AI-TF”) gathered for a retreat.

*A cross-functional team driving business transformation across multiple domains through the use of AI, beyond any singular organization or initiative.

At Mercari, AI tools are now used by 100% of employees, meaning that individual productivity gains are the new normal. However, our transformation does not stop here. Now that individual members have adapted, the next challenge is transforming the organization itself.

In this article, we take a look at where the AI-TF stands today and share a behind-the-scenes look into their energized retreat.

“At this rate, we may not make any progress at all.”

To kick off the retreat, event organizer Shunya Kimura (@kimuras / Vice President, CTO Japan Business) made a blunt admission.

At our current pace, we can’t really say that Mercari is making progress toward becoming AI-Native. That sense of urgency led us to plan this retreat.

It’s true that we’re all now using AI tools. But can we really say our work has shifted to an AI-first model? Ideally, countless agents would handle the tasks suited to automation, while humans focus solely on creative work, final decision-making, and accountability—but we’re not there yet.

And if we continue at our current level of commitment, we may never get there. At this stage in our journey, I want us to realign our understanding and reaffirm our commitment to seeing this transformation through. This is why we’re here today.

Masato Yamamoto (@mark / Executive Officer, SVP of Japan Business), also an organizer of the event, shared this same sense of urgency.

The mission of the task force is to go beyond existing frameworks, step into new domains, and make an impact. At times, that can be a lonely and challenging role.

That’s why it’s important for us to come together, align on our goals, and move forward as one. I believe this will empower us to achieve even greater thingstogether. Both @kimuras and I came here today determined to make as many decisions as possible. Say everything you’ve wanted to say, share everything you’ve wanted to share, and let’s get into it!

And with these passionate words from the leaders, the two-day retreat began.

AI-powered preparation

The defining feature of this retreat was the level of preparation that took place before it began.

Members from all eight domains defined as strategic focus areas—from product development to corporate governance—prepared their documents so thoroughly that no further explanation was necessary. The shared Notion workspace was filled with comments and mentions, so that by the time the retreat began, most key points had already been addressed.

So how did participants digest such a large, highly technical body of information? With AI, of course. Even when content extended beyond their areas of expertise, members used AI to quickly grasp key points and deepen their understanding. The idea was to solve all issues that could be handled by AI before the event, and focus the retreat on discussions that truly required human input.

Meaningful conversations were already underway before the first day, virtually guaranteeing the event’s success before it even started. What followed were two days of intense discussions centered on these refined initiatives and challenges from each team.

Icebreaker bingo tournament

As an icebreaker, members were given bingo cards and went in search of colleagues who matched the descriptions on their cards, such as “someone wearing Mercari merchandise” or “someone whose native language is English.” This activity sparked lively conversations among members who don’t typically interact, energizing the room and setting a collaborative tone for the sessions that followed. 

Presentations across eight domains, and a flood of feedback

Next came the main presentation session. Normally, this kind of session would start with progress reports from the leaders of each domain, but since everything had been thoroughly discussed in advance, detailed explanations were intentionally skipped. Instead, each team shared a concise summary of additional context.

Following each presentation, @mark and @kimuras gave rapid-fire feedback from a management perspective. This helped align insights across domains and sharpen each team’s focus moving forward. For the participants, this was an invaluable learning experience.

Here are some examples of the feedback received:

“In the HR domain, I want you to rethink everything from the ground up with an entirely AI-based model, rather than just extending current approaches. You should question even basic topics like evaluation timing. In the future, evaluations could be continuously updated, with salary adjustments happening weekly. Draft a roadmap that considers even these ‘unthinkable’ scenarios.”

“I really like the concept of resolving all customer support issues within five minutes. But from an AI-Native perspective, we should also be able to prevent issues before they occur. The priorities you established in the previous scope could dramatically change due to AI, so redefine your strategy with that in mind.” 

Deep dives and shared insights

Next, teams moved on to deep dives into specific domains, incorporating the feedback they had just received. The results of these discussions were then shared with the entire retreat, followed by further comments from @mark and @kimuras, creating a rigorous cycle of iteration. Alternating between focused discussions about individual issues and retreat-wide synthesis continually pushed members’ thinking to the limit.

The real issues come into focus

Over the course of the discussions that bridged eight domains, deeper organizational challenges began to surface. One comment from @mark stood out among the domain-specific feedback:

“When a company commits to serious transformation, it will inevitably face resistance. We must be prepared to accept that certain roles may become unnecessary due to AI. This requires the determination to overcome some difficult truths. Even so, our top priority is the goal we’ve set. What changes do we need to make right now to achieve it? Don’t lose sight of what that goal truly means, and keep pushing forward.” 

Some of those “difficult truths” included:

1. We think we’re aligned, but we’re actually not on the same page at all

Even though we talk about breaking down silos across roles, in reality, we often don’t even share a common language with the team next to us. So, the real bottleneck is differences in underlying protocols, not the boundaries between job types dissolving due to AI. For instance, even a basic term like “spec” is interpreted completely differently by product managers, back-end engineers, front-end engineers, and users.

2. Implementation is faster, but decisions are not

The source of bottlenecks has already shifted from technology to people. Even with engineers writing code at lightning speed with AI, we will not be able to launch services any faster if the processes for determining release timing and stakeholder alignment are still sequential.

3. Enthusiasm for an AI-Native transformation has not yet spread across the company

Currently, this excitement is limited to only a few enthusiastic members, and the framework needed to drive company-wide behavioral change is not yet fully in place. While AI-driven discussions are the norm within the AI-TF, pre-AI assumptions still remain in many areas outside the task force.

Each employee’s commitment to follow through

As @mark pointed out, some members were unprepared to accept the pain of sunsetting roles and lacked the resolve to break away from existing frameworks. This is the real, fundamental truth we must face.

True transformation requires more than simply adopting convenient tools—it involves breaking down and rebuilding all processes, from roles and evaluation systems to decision-making processes in executive meetings, without holding anything sacred. This process, and the pain that comes with it, are necessary to face difficult truths.

During the retreat, the AI-TF members witnessed the unwavering commitment of @mark and @kimuras to face these truths head on, and that conviction became a source of energy for all.

At the closing session, participants shared their personal commitments for the days ahead. As with the preparations before the event, everyone wrote their commitments in a shared document, allowing them to be shared in real time. These were not just reflections of past behavior but forward-looking promises to themselves and one another.

To convey the energy of the venue and the determination of each member, here are some excerpts from the document:

“We’ve been too focused on the ‘how’ and not enough on the bigger picture. We need to fully own the question, ‘Can we really do this?’ And to stop asking whether we can do this and start asking whether we will.”  

“Being able to meet many members face-to-face again led to extended conversations and discussions. Going forward, we need to take bold action and rapidly create, test, and scale what works. Our next actions are now clear.”

“Based on the feedback we received, we developed a bolder vision and more ambitious goals. Because of the discussions we had beforehand, we were able to make significant progress in less time. We were also able to align with other teams and their goals.”

“Since we prepared so much in advance, I felt more engaged during this session than at any other we’ve ever had.”

“Some of our most productive moments were the informal conversations we had over dinner. We uncovered ideas about how to use what we already have, ideas for new initiatives, and solutions to issues we previously thought were blockers.” 

“What we discussed has no meaning if it can’t be realized.”

At the end of the retreat, @mark reflected on the significance of the venue:

This venue has been a place where Mercari and Merpay have gathered at key moments for transformation, fostering strong bonds. When we were choosing the venue for the AI-TF retreat, I knew it had to be this place.

There will be challenges when you return to your teams. In those moments, I hope you’ll draw on the experience of coming here together, making a shared commitment, and turn it into energy for tomorrow. What we discussed today has no meaning if it can’t be realized. I hope this retreat becomes a significant turning point for all of us to move forward together.

@kimuras also asserted that this retreat represented a mutual commitment between management and members.

We may have been strict at times, but that’s because we must keep our aim high. You each made a commitment today to see this transformation through, and I hope you hold onto that. I will do the same. As your VP, I will continue to support you fully. And if you feel I’m not fulfilling that role, I want to hear it. I take responsibility for what you’ve declared today. Let’s make this happen together.

The true value of this retreat was in bringing people together who share a vision, creating space for honest dialogue, and rebuilding trust across boundaries. 

These connections started driving collaboration across domains from the very next day, creating a compounding effect. The retreat may be over, but the transformation has only just begun—and signs of real change are already emerging across the organization. 

The cross-functional connections forged at the AI-TF will continue to spread across the organization, shaping a new AI-Native way of working at Mercari through every possible effort.