To shed light on how Mercari’s AI-Native professionals are making the most of AI, we’ve started an interview series called “Yurino’s AI Question Corner,” featuring Yurino Horiuchi (@Yurino)—a Mercari intern who currently works on the AI Task Force and will be joining us as a permanent member in 2026! In volume 2 of this series, @Yurino has a chat with product manager (PM) @jay.kim.

Originally from South Korea, he came to Japan after living in New Zealand and Australia. Having built a career as a data engineer and data analyst, he is now passionate about transforming user experiences as a PM.

In his interview, @jay.kim shares a powerful message: AI is set to become as essential as smartphones. At the same time, he also mentions that the extra leeway we gain by letting AI handle portions of our work actually helps us rediscover what makes us human.

Join us for a thought-provoking conversation about coexistence with AI!

Always looking for where AI fits in at work

@Yurino: So what do you usually use AI for?

@jay.kim: In the previous article, @Yuka said that her work starts by greeting ChatGPT with a cheerful “Good morning!” I’m kind of the same way—whenever I’m typing something, I’m always thinking about how I could restructure what I’m working on if only AI was at the core of the task.

@Yurino: So you’re always thinking about where AI could fit into your work! What specific tools are you using?

@jay.kim: Lately, my personal favorite has been Claude Code—it’s one of the AI tools that Anthropic offers. It’s mainly a tool for engineers, but what I really like is that it creates nicely structured documents with headings and bullet points. That way, I can just copy and paste it into Notion (or have it updated on the MCP server), and all that’s left for me to do is a bit of fine-tuning.

Also, since the voice recognition feature has really improved lately, it’s more efficient for me to speak to the AI model in my strongest language, which is English, rather than typing out everything.

@Yurino: What made you decide to ramp up your use of AI?

@jay.kim: In 2025, I became a father and took childcare leave, returning to work in April. Around that time, news about AI was popping up everywhere, and I was sold—the world was on the cusp of a paradigm shift. In the past, I actually tried to use AI to generate SQL (query language used to work with databases) but due to the model’s hallucinations, it didn’t exactly go smoothly. At first, I was skeptical of AI, but the models just kept getting better and better. By the time I returned to work, I really felt like the technology was finally ready for business use. At Mercari, there was also a movement to ramp up AI usage, so I really appreciated being able to move forward in the same direction as the company.

@Yurino: You really chose a good time to return to work! So, once you were back, how did you go about learning to use AI?

@jay.kim: I wouldn’t say I was studying. It was more about getting a feel for what works and what doesn’t through hands-on experience.

Since there was already a selection of tools available that anyone could use, I just dove in and started experimenting. As things progressed, I tested out the hypotheses that popped into my head—like figuring out what situations these tools actually boosted efficiency for and what tasks were best left to people—and compiled my own playbook of how to make the most of them.

I think it’s important to figure out how AI will work best for you.

AI that protects your focus time with your colleagues

@Yurino: PMs have a wide range of tasks and interact with all sorts of people and teams. Have you had any moments where being a PM made AI especially useful or where AI really made a difference for you in this role?

@jay.kim: First off, my ideal vision for the future of PMs is this: Imagine a meeting where people of various positions and backgrounds come together, and right there on the spot, an LLM instantly organizes everyone’s thoughts and even helps you figure out the best way to get your point across. Wouldn’t that be great? Right now, I have to type out my thoughts for the AI model to sort through, but just imagine how much easier life would be if AI could automatically pick up on what I’m thinking and suggest the best way to get my message across, by also reading the emotional state of the other people present and the vibe in the room! At the same time, I feel that the better a PM gets at using AI, the more important their people skills and ability to communicate become. That’s why I make a conscious effort to separate the things people should handle from the tasks we can leave to AI.

In my work as a PM, AI is turbocharging how quickly I make decisions, especially when it comes to finding information. Mercari is a company that champions Move Fast as one of its core values. While things move at lightning speed, insights often don’t get fully documented. My detective work would start with the classic question, “Who used to handle this project?” I’d ask someone who seemed like they might know, and they pointed me to someone else who might have the scoop… and so it began. Before I knew it, I found that it wasn’t unusual for the person I needed to talk with to have already left the company.

But these days, Notion AI is super useful! You simply have to ask, “So, how did this project turn out?” and it’ll gather information from all your internal tools—Confluence, Notion, Google Drive, Slack, you name it. Finding information has become so much easier.

@Yurino: That’s true! You really become the one keeping the project moving forward and reaching out to all sorts of people along the way.

@jay.kim: Furthermore, I believe it not only increases my efficiency but also shows consideration for other people’s time. When you ask someone for help, you’re actually making them press pause on whatever else they’re working on. But if you ask AI, you won’t have to take up that person’s time. Let AI handle the simple tasks and research, so we can make the most of people’s time. I think that’s the way AI is working best for me right now.

AI usage reveals the value of what only humans can do

@Yurino: How are you using the time you’ve saved?

@jay.kim: I’m using it to come up with better solutions that only humans can think of. Instead of rushing releases, I’ve had more time to dig into the real user issues and the validity of our solutions, which has boosted the quality of our output. In the past, we tended to overlook edge cases when drafting specifications, but now we’re able to cover even error behaviors and display messages, so we can share things in a much more polished state. Also, since we can get AI to perform impact analysis and code review when adding new features, we can now have AI interpret the code directly and resolve issues on the spot, where we used to need to consult with other people. In other words, the time we save is used to enhance quality rather than speed. That’s the mindset I bring to using AI.

@Yurino: We let the LLM pinpoint any blind spots in the specs and keep leveling up the quality as you get feedback. It’s also become easier for you to improve quality on your own.

@jay.kim: That’s right. These days, people are starting to use AI to improve the quality of communication. Since my native language is English, when I write Slack messages in Japanese, the nuances of the language can be tricky for me. That’s why I sometimes ask AI to help me figure out how to communicate in a way that will make everyone feel psychologically safe.

@Yurino: I love that the time saved with AI isn’t just about working faster, but also about improving quality and connecting with people! Are there any tips you keep in mind when using AI, or advice you’d give to beginners?

@jay.kim: First of all, always remember that you can’t lean on AI for everything. Whether you’re working on code or a Slack message, it’s important to be mindful of your work and make the final call to ensure that sharing AI-generated material without making any tweaks will not pose a problem. Additionally, one of the most effective things you can do is avoid becoming complacent. Much like how just about every business professional can now use a smartphone or computer without a second thought, in a few years, not being able to work with AI might mean you won’t even get hired. It’s important to think of AI as an everyday essential on the same level as your smartphone and to get comfortable using it as a natural part of your daily routine. Rather than vowing never to use the technology simply because it didn’t work for you before, keep an open mind and keep trying new things.

@Yurino: You really have to keep using it until it just feels natural. Lastly, do you have any words for those eager to master AI?

@jay.kim: I really liked what Yuka said in the first article about how AI gives us wings. I do feel that the range of things I can do has truly expanded. Allow me to dig a little deeper: In Greek mythology, Icarus flew using wings he made out of wax, but he flew too close to the sun, the wings melted, and he fell into the sea. AI is just like that—it can be a handy and powerful set of wings, but if you use it the wrong way, it can also be dangerous. That’s exactly why you need to take responsibility for anything you create with AI. The walls around job descriptions are starting to topple, allowing anyone to try their hand at anything, and communication skills and leadership matter more than ever. Now that AI can do virtually anything, we must not forget our humanity.

@Yurino: In this age of AI, it’s more important than ever to value the human touch and soft skills—and to take real responsibility for what we create.  I’m glad we had such an interesting conversation that dove into the heart of what work is all about. Thank you so much. I learned a lot!