Dryness, split ends, breakage, and AGA (androgenetic alopecia). Many people struggle with various hair concerns. Bringing innovation to this field of hair care is KNiT Inc.

The company's originally developed AI image analysis technology, "GeXeL," offers a groundbreaking new service that can evaluate the condition of hair from just a single haircut.

The service is very simple to use. When you upload an image of your hair to the AI system, the AI extracts features such as size, shape, and color to analyze the condition of the cuticles on the hair's surface. By quantifying the degree of hair damage, it has become possible to propose appropriate care methods tailored to each individual.

CEO Masataka Kubouchi, after graduating from Tohoku University's Graduate School of Engineering and working for a major chemical manufacturer, founded KNiT in 2023. With the vision of "Reshaping the world's conventions from a single image," he is passionate about creating new value for society. We asked Mr. Kubouchi to tell us about this advanced image analysis technology.

Reducing the Routine Work of Researchers and Engineers

What inspired you to start your own business?

Kubouchi: As a researcher at a chemical manufacturer, I was measuring the number and size of powders one by one, and I always felt it was inconvenient. For example, in (lithium-ion) battery research, it's necessary to check whether foreign substances of different types and characteristics are mixed in with the raw material powder. If foreign matter gets in, there's a risk of short circuits, performance degradation, and in the worst case, fire. It's a very important process directly related to the performance and safety of the battery, but image analysis in R&D is often done by hand, which takes a huge amount of time and tends to be monotonous. The desire to somehow solve this hardship was the catalyst for starting the business.

How did you solve it?

Kubouchi: Listening to those around me, I found that many researchers and engineers had the same problem. My goal was to free them from their daily routine work. That's why I developed "GeXeL," a cloud service for researchers that utilizes AI image analysis (a portmanteau meaning the ability to analyze an 'Image' at the 'Pixel' level). Its core technology is instance segmentation, a technique where improving accuracy is particularly difficult in AI image analysis. It fully automates the quantification of visual information. Image analysis takes only a few minutes. Running costs are also low, and it can reduce about 90% of the simple tasks of researchers in fields like chemistry, food, medicine, and beauty, such as powder analysis, cell analysis, and skin analysis. This allows researchers to devote more time to their primary, creative work.

So, it started as a service for researchers and engineers.

Kubouchi: Yes. While it helps reduce labor and image analysis costs, from a broader perspective, it also contributes to improving product quality through quality control based on objective data. For example, qualitative and ambiguous know-how that previously relied on the experience of skilled workers can now be presented efficiently with quantitative data. It can be applied not only in R&D departments but also in factory quality control.

Why did you start with the beauty industry first?

Kubouchi: It all started when we exhibited at a food trade show. At the show, we explained about the pores in bread and confectionery. When you observe the inside of bread with an electron microscope, you can see gaps called "pores." We've found that the size and number of these pores, and the thickness of the membrane surrounding them, affect the texture we perceive. For example, the more small pores there are, the fluffier and softer the texture. Conversely, fewer, larger pores result in a chewy yet light texture. With "GeXeL," it's possible to numerically evaluate these previously abstract textures through image analysis. Furthermore, by optimizing manufacturing conditions using machine learning, we can explore creating new textures. The ones who showed interest in this technology were from Nakano Seiyaku, a hair care manufacturer who happened to stop by our booth.

That was an unexpected connection.

Kubouchi: Yes. They asked if we could apply it to the analysis of cuticles on the surface of hair. At the time, there was no AI that analyzed the hair's surface, but by using the AI we had accumulated, we were able to achieve high-precision analysis. By linking the cuticle analysis results with data from machine learning, we succeeded in accurately estimating the state of the cuticles just by looking at the hair's surface—something that typically required cross-sectional observation. This allows us to grasp information such as the degree of hair damage and hair age.

So many people have concerns about their hair.

Kubouchi: Women, in particular, have various concerns such as hair volume, dryness, and frizz. However, as a solution, they mostly try shampoos and treatments chosen based on TV commercials or fragrance, without knowing if they are truly suitable for them. We believed that by utilizing cuticle analysis AI, we could propose products that are genuinely right for a person's hair condition. This technology not only solves consumer problems but also leads to improved customer satisfaction and repeat rates for hair salons.

Solving hair problems can boost confidence and bring peace of mind, can't it? What makes your service unique, what are its strengths compared to others?

Kubouchi: Traditionally, hair evaluation was mainly done by hairdressers who would judge hair color, shine, thickness, and damage by sight and touch, without using special equipment. This method is easy and quick, but it's subjective and lacks objectivity. Furthermore, evaluating cuticle damage required special equipment to examine the surface or cross-section of the hair, which was very difficult. In recent years, AI-based hair diagnosis has been introduced to overcome this subjectivity. However, many existing AIs are still trained on subjective data, such as a 5-point visual rating scale, so they haven't fundamentally solved the problem. The decisive difference with our technology is that we provide detailed analysis based on scientific evidence, not just subjective evaluation. By using quantified data, we can more clearly present correlations and causal relationships regarding the hair's condition.

Isn't there a risk of commoditization with just AI technology?

Kubouchi: By combining machine learning with sensory evaluations by hairdressers and hair data, we have established a database that is difficult to imitate.

What kind of hair salons are you providing this service to, and what does the service entail?

Kubouchi: We are providing a beta version of the service at two hair salons associated with Nakano Seiyaku. It's not yet open to the public; we're at the stage where clients of cooperating salons are trying it out, and we're making improvements based on their feedback. We plan to launch around February or March 2026, when the ICT Startup League project concludes, and we hope to officially start the service six months after that.

What has the reaction been from the salons and customers?

Kubouchi: The response has been very positive. From the salons' perspective, they commonly say that being able to visualize the customer's hair condition, which was previously invisible, makes counseling easier. Also, because they can recommend products like shampoos and treatments based on concrete data, they expect in-store sales to increase. We believe this will be a good service that contributes to revenue not just from haircuts but also from product sales. Customers, too, have a need to know the state of their own hair, so we've received high praise from both sides. As for pricing, it will vary by location, so we plan to decide on that while continuing to consider feedback.

How do you want to grow the business in the medium to long term?

Kubouchi: First, we will further develop our hair cuticle analysis. In the future, we want to expand the service to visualize not only hair age and damage but also risk signs of AGA (androgenetic alopecia). Also, since pet fur has cuticles, there's a possibility that we could assess the health of cats and other animals through analysis. We are currently offering a closed service at two salons, but in the future, we believe it can be used not only in hair salons but also in nail salons and beauty spas. By analyzing hair and skin conditions during a customer's waiting time, we can improve the quality of counseling and contribute to in-store sales and branding. Now that we can quantify 'beauty,' which has been subjective until now, we aim to develop a comprehensive beauty analysis service.

Weaving New Value into Society

It seems like you'll be getting calls from outside the beauty industry as well.

Kubouchi: The technology of 'quantifying images' has a universal strength that can be applied to various fields. We believe that by discovering previously unseen correlations and causal relationships from image data, we can create new services and businesses. Our joint research with Nakano Seiyaku began because they took notice of our technology. While some projects start from our own initiatives, as our technology becomes more well-known, we're increasingly receiving various proposals directly from companies. We are hopeful that we can solve problems that were difficult to solve before with the approach of quantifying images, enhance our technological formula, and create new services.

Specifically, in which fields do you foresee business development?

Kubouchi: We receive offers for business partnerships from various fields, but as a startup, we don't think it's wise to disperse our resources. Therefore, we plan to narrow down our business areas based on the keyword "things with universality over time" and focus on fields where we can create a major impact. Currently, we are particularly focused on the following three areas:
① Beauty, Health, and Healthcare: With the advancement of automation by AI, we predict this market will expand further.
② Food and Agriculture: This is a universal theme in human activity, and we want to develop businesses that meet the needs of producers and consumers.
③ Medical Care: This field relates to the very essence of human beings, such as cells and diseases, and also has universal needs.
We want to concentrate our resources on these three areas. Originally, our strength is in chemistry, so knowledge from the chemical field forms the basis of all these businesses.

Through your business, what kind of society do you ultimately want to create?

Kubouchi: Our mission is "to weave unprecedented value into society." The company name "KNiT" represents this philosophy. We aim for the value we create to become firmly established in society, like a knitted fabric, and to become an indispensable presence. The pillar of our business, "quantifying images," reveals previously unseen causal relationships and correlations, pioneering new markets. Our cuticle business visualized evidence that was once tacit knowledge using AI. In this way, our goal is to implement things that were previously unseen into society and "weave" them in so that they become commonplace. We have a strong desire to implement new things in the world, and for that, we value cooperation with partner companies.

What about overseas expansion?

Kubouchi: Currently, our business is mainly domestic, but we are considering overseas expansion in the future. The idea of universality I mentioned earlier applies not only to the axis of time but also to spatial universality, meaning the overseas market, where we believe there is a need. Of course, data on people differs by country and region. Even just with hair quality, tendencies change greatly depending on thickness, waviness, diet, and cultural differences. When expanding overseas, we will need to build a local database from scratch. However, the act of solving hair-related problems is common worldwide. With this universal theme as our axis, we plan to expand our business overseas someday.

You have big dreams. Lastly, what are your expectations for the ICT Startup League?

Kubouchi: I want to find new partners and collaborators in the cuticle field we are currently working on. In particular, B2C product development and media strategy are our weaknesses, so we are looking for people who can cooperate in those areas. We also have expectations for fundraising. We are considering raising funds in a Series A round, so we would be grateful for support in connecting with VCs (venture capitals).