xMap, an AI that understands geographical space, instantly answers questions when you circle an area of interest on a map, and is attacking the global market from Tokyo
There is a theory that San Francisco has become a turf for AI startups, but Tokyo should not be underestimated. The interview with this AI startup was conducted in a co-working space in Tokyo, and in the next room, the team from Sakana AI, famous for being the fastest unicorn in Japan since its founding, was working. If there are legends under the same roof, other startups may be inspired. They say that "the environment makes the person," and the same is true for business.
xMap is not yet a unicorn, but it is a startup that has grown rapidly in just under a year since its founding. Founder and CEO Mo Batran is originally from Egypt, but he specializes in spatial information science and came to Japan to study at the University of Tokyo. He then co-founded LocationMind, which develops location-related AI and space businesses, and served as business development director before launching xMap last year. He posted about the service on LinkedIn every day , and one day, suddenly, a post was created that was viewed more than 1 million times. This post triggered a flood of inquiries from many companies. The company won contracts one after another with major companies, including Coca-Cola, and achieved sales of 600,000 USD.
What is an xMap?
xMap's core service is a geographic information analysis tool that utilizes AI. Users can specify any area (polygon) on a map and ask the AI various questions about that area. The AI can instantly provide answers to questions such as, "What is the population of this area?", "How many gas stations are there within a 1km radius?", and "How many people are there at breakfast, lunch, and dinner?" Batran said, "Until now, answering these questions required time to collect and analyze data, but xMap makes it possible to do so instantly."
This is made possible by the data we collect, our proprietary models, and technology that uses large-scale language models (LLMs) to understand the context of location information.
A typical LLM doesn't understand the concept of location. For example, if you ask, "How many restaurants are there here?" it doesn't know what "here" means," says Batran.
xMap makes this feat possible by training its AI models with a combination of open government data, data purchased by xMap, and data crawled from the web. Batran declined to reveal specific data sources, citing confidentiality, but xMap currently has coverage for Tokyo, New York City, and Riyadh. Riyadh's coverage is likely related to the fact that many of xMap's users are investment companies, one of which is the Saudi Arabian sovereign wealth fund.
Expanding possibilities
In Japan, there are real estate services that allow you to click on a house on a map and instantly see the selling price, buying price, rent, etc., but xMap's scope is much broader than those. It implements functions such as contextual understanding of location information, multimodal learning, continuous learning, and data freshness maintenance, allowing for more flexible and intuitive analysis than traditional GIS tools.
xMap is currently used in a wide range of industries, including real estate, retail, consulting, and investment companies. In the real estate industry, it is used for property evaluation and development planning, as it can instantly grasp the demographic trends and distribution of facilities around a property. In the retail industry, it is used for location analysis when opening a new store. Investors and asset management companies analyze changes in demographics and economic indicators in specific areas, which is said to be useful for investment decisions. It is also beginning to be used in urban planning and public policy formulation, and is attracting attention not only from private companies but also from the public sector.
xMap is currently available in three cities, but the company is looking to expand globally. The xMap team currently consists of 10 people, of which the four founding members, including Batran, are all from Africa, with other members from Europe, Pakistan, and India, as well as Saudi Arabia and Nigeria.
xMap recently raised funds (pre-seed round according to Crunchbase ) from Japan-based Shizen Capital , and is aiming to raise a total of US$2 million in this round. "Startups based in Japan that aim for the global market are still rare," said Batran. We asked him what kind of investors he would like to raise funds from in the future.
In the Japanese market, startups tend to be valued at low valuations. Many investors see startups as simply system development companies and don't fully appreciate the potential for global growth. We also approach investors globally. We are looking for investors who have a global perspective and can support the growth and virality of our products. (Batran)
The pain this startup is trying to solve and the solution it offers are really easy to understand. Seeing is believing. Except for cases where the technology itself is unique, services that take time to explain or that we, the potential users, cannot easily understand rarely become a big hit. In that respect, xMap's explanation does not take the one minute that is considered the standard for an elevator pitch, and if 10 people see the service, 10 people will be able to understand it. With the rise of generative AI, we may see more startups like this in the future.