I tried something like this.
'Tell me a manga I should read on a tough night.'
The answer from the AI assistant 'Gemini' was gentler and more accurate than I imagined. If it's just one volume for tonight, Keigo Shinzo's Hirayasumi; if you want to reframe your hardship over a long timeline, Kanehito Yamada and Tsukasa Abe's Frieren: Beyond Journey's End; if you want to be pushed forward by someone slowly recovering, Chica Umino's March Comes in Like a Lion. Three works were listed according to the mood.
'It feels like someone sitting next to you without telling you to "do your best,"' 'The flow of time over decades quietly affirms loss,' 'A lonely young shogi player is saved by the kindness of three sisters across the river.' For each, the characteristics of the work and 'why it works for this night' were carefully added.
It wasn't by genre, volume count, or ranking. A manga shelf rearranged by 'mood' and 'scene' already existed within the AI.
I am the CEO of Comici, a manga DX startup. I face manga and publishers every day, yet I don't have the confidence to answer so clearly and organized when asked, 'Tell me a manga for a tough night.'
AI is already entering a stage where it recommends manga based on life scenes and emotional states, transcending genres and rankings. What interests me is what lies ahead. What exactly is the AI looking at when it chooses manga? And is the manga industry prepared to continue responding to that inquiry over the long term? What is actually happening behind the scenes where AI seems to be answering so cleverly on its own?
I wanted to think about this a bit more carefully.
Why Walmart Arranged Frozen Foods by 'Breakfast' and 'Lunch'
Let's step away from manga for a moment and talk about US retail.
Walmart has begun rolling out a new store concept called 'Store of the Future' in Texas and California. What's particularly interesting is the story of the frozen food section.
A surprisingly simple change was introduced. The frozen food section is categorized not by product category (pizza, frozen pasta, frozen bowls) but by the timeline of daily life: 'Breakfast' and 'Lunch.'
Pizza goes to the pizza section. Frozen pasta goes to the frozen pasta section. That was the traditional frozen food aisle. In Walmart's new stores, the shelves are divided by the life timeline of 'Breakfast' and 'Lunch.'
Technically, they aren't doing anything amazing. However, the meaning of this classification is significant. When an AI answers the question 'Are there any easy and healthy frozen foods for breakfast?', it doesn't refer to the product name itself, but to 'contextual data'—what kind of life scene that product appears in and what kind of person chooses it.
With only the label 'This is a pizza,' AI cannot answer the question 'Is there a good frozen food for breakfast?' Because people talk to AI in natural language, AI can only find it if there is a meaning attached like 'This belongs to a morning life scene.' The location of the shelf where the product is placed itself becomes the entry point for data that records that meaning.
Walmart is a company that has 'Everyday Low Price' in its DNA. The only reason such a company would go through the trouble of editing its sales floor is that it is rethinking 'how to make its DNA function in the AI era.'
Actually, Manga Magazines Were the 'Data Rooms'
Now, back to manga.
Imagine the manga section of a bookstore. Shonen manga, Shojo manga, Seinen manga. Or by publisher, by author, by volume order. Shelves are divided by genre or magazine name, and you go looking for a specific author or series. That is the general arrangement.
What if you suddenly asked in front of that shelf, 'Is there a manga I should read on a tough night?' Neither the Shonen manga shelf nor the Shojo manga shelf will answer that question. It's the exact same structure as the frozen food aisle.
Unless meanings like 'perfect for a sleepless night,' 'gives you a push on a rainy morning,' or 'something to immerse yourself in during a long-awaited holiday' are layered on top of labels like genre, author, or magazine name, even AI won't be able to find them (of course, some bookstores do create manga shelves based on themes).
So, who does that 'meaning-making'?
Let me talk about a slightly different world. I came across an interesting analogy while reading an article explaining how AI is used in modern military contexts.
There is a platform that organizes satellite images, communication records, and various databases coming in from all over the world into a network of relationships, such as 'this person belongs to this organization, is in this location, and appears in this communication.' This is a service by a company called Palantir. Then, the AI on top of it makes inferences from the organized data and writes reports.
One explanation described the relationship between the two like this: The platform is like an 'editorial department's data room and proofreading system.' It is the foundation for organizing, linking, and managing access rights to information gathered from around the world. On the other hand, AI is like a 'highly capable external analyst who doesn't know the internal affairs of the company.' It reads the data provided by the data room, finds patterns, considers scenarios, and writes reports. However, the AI cannot access information that the data room does not provide.
I nodded strongly at this analogy.
AI is an excellent analyst. But it is an outsider who doesn't know the internal affairs. In a place without a data room, AI is just a smart person with general knowledge. Conversely, in a place with an organized data room, AI behaves like an expert in that field.
What happens when we overlay this onto the manga industry? The reason AI can return mood-based answers to the question 'Is there a manga I should read on a tough night?' is that there is a data room where someone has previously assigned meaning to and organized the manga. Without a data room, AI can only give general knowledge-level answers.
I believe that the role of that data room in manga has long been played by the editorial departments of manga magazines.
A manga magazine is not just a medium that collects works. It is the very act of assigning meaning: 'The group of works in this magazine reaches this kind of reader in this kind of mood.' By picking up a magazine, readers were able to find a group of works that fit their mood without even knowing it. The name of the magazine itself functioned as a shelf of manga with assigned meanings.
At this point, we can see what the manga industry needs now. It is to re-assign meanings like 'effective for a tough night,' 'resonates on the night before graduation,' or 'refreshing after an all-nighter' as data to each individual work, in units even finer than manga magazines. It is to use the power of data to enrich the sensibilities that manga magazine editorial departments have carefully refined over many years.
What Comici Wants to Do is Something Much Simpler
What we at Comici are trying to do overlaps exactly with this.
Until now, data surrounding manga works—publishers, digital bookstores, SNS, anime, merchandise—has existed separately across the industry. Completion rates, PVs, billing, SNS reactions, fandom enthusiasm. All of these are important materials for talking about the value of manga, but they have hardly been evaluated across the industry.
Comici is building a foundation that bundles this data and gives a contour of 'meaning' to each manga work. It is a data foundation that serves as a proper basis for decision-making regarding the development of works, such as whether to continue serialization, move toward a screen adaptation, release products, or expand overseas.
But what I really want to do is something much simpler.
To create a state where the manga industry as a whole can respond to questions like 'manga to read on a tough night' or 'manga that gives you energy on a Monday morning.' To use the power of data to enrich the sensibilities that manga magazine editorial departments have cherished for many years.
I believe that this is the way to increase the number of manga chosen by AI and is also a condition for delivering Japanese manga to readers around the world.
Whether we can offer a reader standing in front of a shelf, 'This is your manga for tonight.' I want to rethink the work at the very root of the manga industry: handing a work to a reader.
