See the original story in Japanese.
IBM Japan held the Demo Day event for their BlueHub open innovation program last week. The latest batch features startups focused on inbound travel businesses.
A number of companies offering inbound travel services participated, with four big corporates: NTT Docomo, Zenrin, Zenrin DataCom, and Softbank, and eight startups: Andeco, Crea Japan, Fesbase, Pretia, Mybase, Metro Engines, Rich Table, and Realista.
Presented services and companies were:
- Fesbase: chatbot platform focused on serving Chinese visitors to Japan (by Crea Japan)
- Quippy for Restaurants: Marketing and product development support tool for restaurants (by Rich Table)
- Sakevel: Sake brewery tourism and personal sake sommelier service (by Mybase)
- SnapGo: AR(augmented reality)-based navigation service (by Pretia)
- Miccossy: Mobile app curating Japanese local festivals (by Andeco)
- Travel experience sharing platform (by NTT Docomo)
- 1Minute Japan: Video-based service for helping foreign visitors to Japan solve problems (by Realista)
- Metro Engines: AI (artificial intelligence) tool that helps hotels determine competitive pricing
In late February, teams composed of a mix of the companies carried out kick-off meetings and held six workshops throughout the three month planning phase. Following the Demo Day they will begin to work on firm plans for commercialization. In the spirit of full disclosure, I was asked to participate in this event as a judge. The following are introductions of the seven services that gave presentations.
Fesbase: chatbot platform focused on serving Chinese visitors to Japan (by Crea Japan)
When planning vacations overseas you may often rely on TripAdvisor and Yelp, but it is hard to tell the atmosphere of a place and whether it is suitable for kids or not. It is easy to imagine the same situation occurring during the expected boom in inbound tourism to Japan. Many tourists from China come to Japan for the cuisine. They may be left wondering how to make a reservation in such situations. That’s where the Yoyaku app uses chatbot and AI to come to the rescue.
Users select the category of restaurant they want to reserve and input details like the area they hope to go to. The app supports voice input so it is easy for travelers to use, and recommended information is delivered leveraging IBM Watson.
Since Fesbase already has an existing project that offers a human concierge, there is a large amount of training data, so it is possible to provide accurate recommendations. If, due to the contents of the question, the bot cannot answer, it is possible for AI to sort and select a human concierge who can. They are currently in negotiations with the Chinese media and also looking into acquiring users before their arrival in Japan.
In terms of business, they are thinking to charge a fee to restaurants based on driving user traffic to them, and in the future, once the number of users increases, they are planning to offer a platform for restaurant searching services to businesses.
Quippy for Restaurants: Marketing and product development support tool for restaurants (by Rich Table)
Quippy provides a restaraunt searching tool on Instagram. It offers users with the chance to discover (even the unexpected) restaurants via searching. When the user launches the app, information linked to the location information of the restaurant and the picture pulled from Instagram is displayed. Since the app will learn what content the user has browsed, it will recommend additional restaurants later on.
For the service geared at restaurants, the tool will show them posts in English on social network sites and restaurant review sites as negative or positive. If a specific dish receives a negative review, it can recommend Instragram examples from another shop to the restaurant to improve. The searching service is free, while the service for restaurants will be offered based on a monthly subscription model.
Sakevel: Sake brewery tourism and personal sake sommelier service (by Mybase)
88 of the 100 foreigners questioned at Narita International Airport said they want to drink Japanese sake. On the flipside, they don’t really have any information about sake breweries. Moreover, more than 99% of such breweries are mid-sized or family run so they may not be able to fully support foreigners who show up unannounced, but even before that 60% of consumption by foreigners visiting Japan is in the Tokyo metropolitan area. The idea behind Sakevel is to utilize tourism resources to awaken the small sake breweries that sleep in unknown areas, thus promoting regional development.
As for the sake itself, the Sakevel app recognizes the images on labels and provides that information in multiple languages. It also introduces personalized sake recommendations for tourists based on information gathered from past posts on social media sites. Since breweries cannot attract customers alone, Mybase plans to offer sake brewery tours along the same lines as a winery tour. Services are set to begin in October this year.
SnapGo: AR-based navigation service (by Pretia)
SnapGo will take you where you want to go without GPS by specifying your location with a mobile picture. This solves the problem of the numbers of foreign tourists who get lost due to lack of multilingual signage in Japan.
The company receives a photo of the place from its premise owner, creates an algorithm from it, and makes it possible to identify places from photographs taken by users through machine learning. As a result, users without GPS on their devices or who experience language difficulties can get to their destination by simply taking pictures.
In order to develop the algorithm it appears they need about forty 360-degree images. Additionally, it is assumed that the destination is input using images or text.
Miccossy: Mobile app curating Japanese local festivals (by Andeco)
Simply put, it is a festival information service for foreign tourists in Japan. It is difficult for users to acquire real-time festival information even if they refer to guide books, etc. People outside of Japan search for “matsuri (literally festivals in Japanese)” on YouTube as much as “Mt. Fuji” so there is a demand. Moreover, the company expects that there are 310,000 festivals all over Japan.
The Micossy app displays festival information in the surrounding area and provides related information like how to participate, and so on, in multiple languages. Their business model is the sale of “happi” or traditional clothing to be worn at the festival. Initially they plan to start by offering information on 100 festivals.
1Minute Japan: Video-based service for helping foreign visitors to Japan solve problems (by Realista)
1Minute Japan is designed for travelers to Japan who come on their own, not as a tour. It provides help to travelers before departure on topics like SIM cards and Wifi in Japan, and once they begin their journey, the company offers transportation and ticket information, as well as a restaurant reservation service. Using Watson to analyze the information learned from these tasks, it can analyze sentiments and classify what kind of problems people in various countries experience. Then they release video content to solve these problems.
The service can be used in this way: the example that, “an American traveler in their 20’s doesn’t know how to eat at a conveyor belt sushi restaurant”, guides them to create video content to help this person. They started test marketing in April and have acquired around 5,000 followers on their Facebook page.
Metro Engines: AI tool that helps hotels determine competitive pricing
Metro Engines can optimize revenue for hotels and inns. The tool targets the 35,000 facilities throughout Japan, with 1.2 million rooms. Many of the rooms were priced using an individual’s expertise. Some chains use tools, but set their prices based on past performance and competing room prices. However, it can be said that this method will become ineffective as the supply-demand balance collapses due to access to private residences, etc.
That is where Metro Engines comes in. It predicts the reservation behavior of guests, and then set room prices through data analysis peppered with budget information, reviews, and furthermore, information on renting private homes and rooms, as well as guest room photos. In addition, It has also visualized human behavior based on mobile behavior data from NTT Docomo, Zenrin, and Softbank and included it.
Based on the big data obtained through these measures, they can tangibly present how a room should be priced. The service was announced on the 25th of last month and is currently under evaluation at 50 facilities. When they link up with the individual systems of hotels and inns, they can verify that the price setting was appropriate.
Watson’s unique open innovation program and future issues
Followed by the previous batch focused on automotive and healthcare verticals, this was the second edition of the IBM BlueHub open innovation program. Prevent, one of the grads from the previous program batch last year, is a startup spun off from Nagoya University Graduate School of Medicine. The team presented a solution preventing lifestyle diseases, which was collaboratively developed with a health insurance company.
Based on his research experience having succeeded in lowering the recurrence rate of high blood pressure, diabetes, cerebrovascular disease and other symptoms to less than 30%, he has developed an online consultation service. He claims that they will aim to gain the consultation efficiency by implementing the Watson technology into a chat-based consultation service.
In the latest batch focused on the inbound travel topic, participating companies were trying to create new values by combining with the unique strength of the services, technologies and resources that they have had.
Metro Engines was highly evaluated because of not only telling hotels optimized room pricing leveraging big data analysis but also predicting details and suggesting ideas about how to improve. The team is outstanding with having already 50 companies as potential clients.
Regarding aforementioned Predict and Metro Engines, they could unveil clear results through the program, however, the ideas of other teams were still “small” in the potential about their technologies and services or their ideas are still rough and unclear in the potential merit to come out of a collaboration work.
Moreover, we saw some projects which had probably managed to adjust their ideas into Watson in the program, meaning that they have turned a means into a purpose there.
It might sound good if we say a collaboration between resourceful corporates and game-changing startups with technologies. However, if the collaboration become a purpose, I assume that it will not bring a good result because of the gap in expectations and cultures on both sides, or the program host’s egotism.
Where is the problem to solve? Does it require any technology? What kind of businesses should be get together to speed it up? Open Innovation is now trending in Japan, that’s especially why I think these well-considered preparation may be more important.
Translated by Amanda Imasaka
Edited by Masaru Ikeda