Fujitsu’s AI story rises in the East
The 2020 Olympics in Tokyo will not just showcase the best athletic talent on Earth, it is also aiming to show off the bleeding edge in digital technology.
Driverless taxis are currently being trialled for the event, while contactless tickets, facial recognition technology to welcome spectators and an artificial meteor shower are all being lined up.
One of the companies looking to display its own capabilities on the Olympic stage will be Fujitsu. For example, the company is working with the University of Tokyo Hospital to build an automated voice translation system to help doctors cope with a spike in the number of foreign patients, which is expected to double to 800,000 in the next three years.
This week, Fujitsu ran its annual Forum event in Tokyo to highlight its new products and services to clients and prospects, and much of the focus was on the company’s four strategic growth areas: AI; IoT; security; and cloud.
The company’s cloud activities are playing a leading role in driving the global standardization of the group portfolio – an area where Fujitsu has historically struggled, but has made progress in the last 18 months. The MetaArc digital management platform and K5 IaaS and PaaS offerings have recently launched in UK, Germany and Finland with the US coming up in the current year.
Fujitsu’s IoT business grew by 40% in its most recent fiscal year, with the largest part of its revenue coming from supporting connected vehicle initiatives. One area where the company is successfully building repeatable IoT platform plays in its domestic market is in agriculture (not the sexiest use case for IoT, but one where it is arguably delivering the greatest returns), and it is having some early success in extending this in markets such as Finland.
In cyber security, the company is seeing strong demand for incident response services – even before the WannaCry attacks landed. Fujitsu showcased a new forensics offering this week, which is aimed at speeding up the time it takes to identify the extent of damage from an attack from several weeks down to under an hour. The company also announced a new tie-up with US-based Dtex Systems to leverage the latter’s user behavior intelligence technology that identifies suspicious activity that could lead to information leaks.
This brings us to AI., and it was the main focus of much of this week’s event. Fujitsu has a 30-year history in exploring the possibilities of machine learning, and has more than 200 AI patents under its belt. At a conceptual level, Fujitsu’s AI service leader Hirotaka Hara believes that because AI’s inability to grasp the complexity of long sentences and its innate lack of common sense will limit its impact on the workforce, but it will find a place in delivering the analytical insight that will enhance the creativity of human workers.
Fujitsu’s proposition is a work in progress, following the launch of the Zinrai platform and brand in 2015, but the general positioning is to create a broad portfolio of standalone and embedded offers, including industry-specific APIs. Naturally, Fujitsu develops the underpinning hardware platform (which can be delivered in on-premise, hybrid or public cloud flavours), and it expects to make some significant performance advances in the medium term through its research and investment in quantum computing architecture.
Fujitsu currently has a core team of 700 specialist AI engineers (a figure expected to double to 1,500 during the current financial year). The company has a pipeline of 500 projects/prospects as of May 2017, and it also offers an interesting breakdown of what use cases are driving the business. Some 20% of its engagements lie in supporting contact centres, applying machine learning to historical data to drive automated voice conversations. Funnily enough, part of the driver behind this growth has been IBM’s aggressive push of its own machine learning-driven call centre service in Japan. Other significant areas of activity include manufacturing and maintenance (17%) and financial services, with Nomura Securities using Zinrai machine learning to detect irregular transactions.
Of course, Fujitsu has some pretty stiff competition in the AI arms race, from traditional players such as IBM, through to Google, not to mention a growing wave of start-ups. Beyond the pursuit of greater computing power, one of the key battlegrounds looks likely to be the development of extensive data sets, that enable the vendors to turn up at their clients with the strongest possible starting point in helping them to understand customer behaviour and operational performance in a wider context.
IBM’s purchase of The Weather Company last year suggested that this is the way that the market is heading, but Fujitsu is not planning similar moves as part of its M&A strategy. Instead, the company has said that many of its customers are looking to monetize their own data sets, and that it sees potential in building joint offerings to take to the wider market.
Fujitsu has talked abut “human-centric innovation” for several years, and while businesses will initially look to leverage AI to either enhance their customer understanding and interaction, or to increase productivity (and Fujtsu is on its own internal journey to automate many areas of its business), it also carries the potential for deliver some huge social benefits. For example, Fujitsu is the first company that PAC has seen present its AI offering in terms of how it addresses some of the United Nation’s 17 sustainable development goals.
The company is working with the Hospital Clinico San Carlos in Madrid to help senior doctors to accelerate the time it takes to identify individuals at risk from substance abuse, depression or suicide. Fujitsu has developed an AI tool with has learned from 36,000 clinical records as well as one million open data sources (such as research papers). The tool can now generate a diagnosis in a matter of seconds that has been found to be 95% as accurate as the conclusion arrived at by eight senior practitioners, freeing them up to spend more time on patient consultation.
Fujitsu is also playing an interesting role in tackling Japan’s inverting population pyramid, with four in ten of the country’s population expected to be over the age of 65 by 2050. It is part of an initiative connecting 7,000 health and social care agencies including hospitals, clinics and pharmaceutical companies with the aim of providing proactive services to support the well-being of patients, taking the focus away from reactive treatment.
In all four areas, Fujitsu has some really interesting capabilities and client stories that tackle universal business and social issues. However, it faces the same challenge as its peers in turning them into repeatable, standard offers to drive both top and bottom line growth at an international level. This is not something that comes naturally for Fujitsu, whose stronghold in its domestic market has been historically driven by the empowerment of the field engineers to lead what often end up as heavily customised system integration projects. However, some important changes are being made. The Fujitsu Labs R&D engine has active projects in 15 markets, the company is ramping up its engagement with partners in Silicon Valley, and Germany is emerging as an important centre of gravity for its IoT activities.
Fujitsu has a distance to travel before it can replicate the kind of board-level relationships it has with domestic corporate giants such as Toyota outside of its home market. But it at least looks to be on the right path.