Cut through AI noise with weekly insights from Japan to AI-proof your career or business. We aim to bring you insights that are: digestible, actionable, and unique (breaking through the information firewall in Japan).

In this week’s newsletter:

It began in the ancient times of September 2025, when ChatGPT launched their “Checkout” feature. This allowed users to complete a purchase directly inside the chat interface, without being redirected to a traditional e-commerce site. Discovery, comparison, decision-making, and payment.. all coordinated by your AI, without leaving the chat window. The equivalent of being too lazy to get off your couch to even go to the kitchen.

AI will discover, compare, help you decide, help you trust, and help you fork over the dough. Its a huge change from the flow we all know and love/hate: search → scroll items → view item → add to cart → checkout.

As has become the norm, Google is not to be outdone for long in the AI stakes. They came along this week and announced a new open standard called Universal Commerce Protocol (UCP), co-developed with major retailers and platforms including Shopify (plus names like Walmart/Target/Etsy/Wayfair per reporting). UCP defines a “common language” for AI agents and commerce backends to coordinate the full shopping journey, from discovery to checkout to post-support. They are explicitly tying UCP to checkout within Search AI Mode and Gemini, which means huge distribution if users accept it.

Okay, so what are the signals to pay attention to here?

Traditional e-commerce sites were about click-through rates, SEO and paid acquisition.

AI-native checkout is more about understanding the users‘s intent with context about their preferences, and building trust in the back and forth between AI and user. We aren’t browsing and shopping any more, we’ll be asking our agent to understand us well enough to handle shopping for us.

If AI handles the purchase intent and transaction in its own window, it makes sense that its no longer just a tool but actually a gatekeeper of demand. This will unlock new revenue streams for the AI companies through potentially transaction fees (?) or affiliatte margins. They’ll also know users better than any search engine or ad company could, through conversations. And once users trust their AI, switching costs will rise sharply.

It all means that the big AI companies will start to control “what gets recommended” and therefore influence demand.

There is the upside of pre-qualified customers.. meaning that customers that show up to the retailers site will have high intent to purchase. But in reality, that might not actually be customers anymore, and rather just AI agents on customer's’ behalf.

And so retailers will lose brand awareness and perhaps just become suppliers to AI agents rather than actually have strong customer relationships. If that happens, expect worsening margin pressures, and dependence on the AI recommendations.

This shopping experience will prioritize “satisfaction” and “convenience” and not “exploration”. But for many people, exploration is a part of the deal. I know lots of annoying “travel hacker” types who won’t stop bragging about deals they found. If AI took over this part of their lives, they would become hollow shells of human beings. Lol.

Also, you may trust the AI but incentives become opaque.. you may start to wonder, is this the best product for me? Or for the platform? And you’ll also have less choice visibility because AI will always summarize and condense down into your “top 3 or top 5” options.

This could lead to a "digital sameness." If an AI agent always picks the "most logical" choice based on structured data, we will lose the serendipity of shopping.. like accidentally discovering an egg maker on amazon that only boils a single egg, while scrolling at 2 am.

All in all, it may not be everyone’s cup of tea, and a resurgent movement in hyper human, no AI shopping experiences may also emerge.

ChatGPT Checkout and Google UCP are definitely signalling a move toward a parallel commerce stack, optimized for agents, not humans. Platforms will compete to own user intent, retailers will be in a fight to the death to be recommended by AI, and customers will trade choice abundance for decision efficiency.. and potentially hate it.

For example, I used to often use this flight tracking/watching tool called Hopper. I think I even paid for it. I used to set alerts for routes that I wanted to keep an eye on (especially back when I was more strapped for cash), and Hopper would let me know the right time to buy the ticket.

This type of rich monitoring application is ripe for disruption by vibe coding, in my opinion. I still want to watch routes from Japan to elsewhere in Asia. So I made an equivalent of Hopper (or at least the flight tracking feature), using Google’s AI Studio.

This was an early version, which took just 10 or 15 minutes.

After this checkpoint, I continued iterating, added Gemini API keys, made it pull real time data, and it actually worked! I am sure if I spent a little bit of admin time stabilizing the app etc. it would be even better. Perfectly serviceable for this use case, and no need to install complicate flight tracker apps.

Highly recommend playing around with Google AI Studio and their Flash model. Building mini apps has never been easier or cheaper!

The Tool: Omakase.ai

What it is: Omakase is a digital "personal shopper" for e-commerce stores (specifically Shopify stores). It’s a multi-modal AI agent that combines voice, chat, and visual browsing to guide customers through a store’s catalog. Instead of a static chatbot that just answers FAQs, Omakase acts as a proactive sales rep, listening to what you might want and suggesting the exact products that fit your intent. As they claim on their site, “chatbots suck”. And they truly do.

Who made it: The tool is powered by Zeals, a very successful Japanese startup founded in Tokyo in 2014 by Masahiro Shimizu. They are widely recognized as the pioneer of “Chat Commerce” in Japan, particularly through their integrations with LINE.

The Name "Omakase": Their product name, Omakase.ai, is a nod to the Japanese culinary tradition of "leaving it to the chef," which captures their goal of letting the AI curate the shopping experience. You just show up to the store front and, “makasete”.

Why you should try it: A lot of people like me haaaate scrolling through pages and pages of ecommerce search results. Omakase can save you the hassle. For store owners, it’s a way to offer a "concierge" experience 24/7 without the overhead. And for us shoppers, its step one into the future of shopping: you don't navigate a website any more, you have a conversation with it. Of course, as we talk about above, will Omakase lose share to the foundational models as the conversation goes up a level to where the customer never reaches the Shopify site? Or will AI (ChatGPT) talk to AI (Omakase) to get the information? A space to watch.

​​This full-day, hands-on workshop shows how generative AI can accelerate early-stage entrepreneurship by taking participants from idea to business prototype in one session. Attendees will work through structured modules on:

The format will prioritize practical application over theory, with AI experts guiding development and strategy work. It targets people comfortable with basic AI tools but new to building real products.. especially non-technical founders and aspiring entrepreneurs. Go check it out!

Organizations involved included Women In Tech Japan (empowering women and girls in STEM) and Tokyo AI (TAI) (Tokyo’s largest AI community).

It has to be the Grok debacle.

Just this week, Elon Musk's platform X was forced to disable image generation features for its chatbot, Grok. The tool's lack of guardrails allowed users to generate non-consensual and sexualized deepfakes of people with simple prompts, triggering immediate legal threats globally. They called it “spicy mode” but its really exploded in controversy and quite a few governments are actually blocking Grok altogether.
It calls into question how long the LLM companies can avoid responsibility for the negative consequences of their products.