The Watson IoT unit will be responsible for managing new solutions and developing IBM partnerships around IoT. One of the reasons for IBM choosing Munich may have been its role in the German motor industry and because it is a major technology hub. This will make it easier for IBM to gain access to the emerging automated vehicle market in Germany.
One of the big goals for the Watson IoT unit will be adding cognitive computing to IoT. This will play well inside the automated vehicle space where companies are looking for greater intelligence inside the vehicles. This goal is to move them away from being the equivalent of slot cars where they not only take information from a wider number of sensors around the road and in the vehicle to being able to make decisions for themselves.
A focus point for European Watson development
The office in Munich is also the first European Watson innovation centre. IBM has said it will be basing over 1,000 staff at the office. What is not clear is how many of these will be new hires and how many will be reallocated from other IBM offices. What is not clear is exactly how the numbers will breakdown. It would be nice to think that IBM is going to create more developer and researcher jobs than consultants and sales people.
The innovation lab certainly hints at a more technical than sales focus and this is good news for Munich. It has lost a number of jobs in the IT sector over recent years as companies have reduced their employee count in difficult market conditions. It will also be interesting to see how much engagement there will be with the German start-up industry. Munich is one of several start-up hotspots in Germany although it is much smaller than Berlin.
There is real scope for IBM to put aside some of the office space to attract start-ups into its own building and therefore rapidly expand the number of applications taking advantage of Watson. However this would be a medium-term goal if it happens at all. The first phase of engagement outside of IBM will be with existing partners and IBM clients who want to do some testing of their applications.
Blurring the edge of IoT and cognitive
One of the key challenges for the Watson IoT unit is how to bring IoT and cognitive together. In terms of industrial sensors, the challenge is pretty simple and ends up as little more than an expert system controlling an oil refinery, chemical works or some other sort of installation.
What IBM wants is to go much further. This means looking at other forms of IoT around healthcare or fraud for example. Once again there are challenges here for IBM. It has long resisted the Dr Watson moniker for its work in the healthcare and health sciences market. Instead it has worked hard to portray Watson as an able assistant rather than a decision maker.
With healthcare IoT it can go further than just ingesting a lot of old manuals. For example it could monitor an entire hospital using data from everything such as CCTV, lighting, heating, medical equipment, water consumption in toilets, basically anything in which you can embed a sensor and gather data. It could then use that data to identify risks ranging from a lack of beds or medical supplies to a slight increase in deaths or secondary infections that indicate a member of staff or visitor is a carrier of some disease.
In the world of fraud IBM has already struck deals with insurance companies to monitor transactions to look for fraud but what about using it for investigatory powers. A claim for personal injury around a traffic accident could see Watson collecting data from weather stations, other vehicles, local CCTV and even the vehicle the claimant was in. From that data Watson could work out the likelihood of the claim being due to driver error, equipment failure or a third-party action.
At the moment, where IBM will go with energising the start-up market to open up this new era of cognitive IoT is open to question. For example there will need to be a conversation around privacy when you think about the personal injury case above. In addition, many of the start-ups will lack access to the necessary wider systems to build their cases and this is where Watson through SoftLayer becomes a key tool.
There are many areas of research where Watson is currently being trialled and in some of these it would be easy to see how the use of IoT to gather greater information could be used. What we really need is IBM to start opening up and talking about how wide a net they will cast, what resources they will provide to start-ups and how it sees cognitive IoT impact lives and legislation.
IBM has already announced it will invest $3 billion to look at the convergence of IoT and cognitive systems. How much of that will be available to the Watson IoT unit to spend in Europe and how the funds will be accessed by people outside IBM needs to be clarified. Hopefully IBM will hook to the surge of start-up competitions around Europe and invest in helping bring things to market.
In the press release Harriet Green, general manager, Watson IoT and Education was quoted as saying: “The Internet of Things will soon be the largest single source of data on the planet, yet almost 90 percent of that data is never acted upon. With its unique abilities to sense, reason and learn, Watson opens the door for enterprises, governments and individuals to finally harness this real-time data, compare it with historical data sets and deep reservoirs of accumulated knowledge, and then find unexpected correlations that generate new insights to benefit business and society alike.”
New Watson API Services announced
To kick everything off IBM is announcing four new families of Watson API Services focused on the IBM Watson IoT Analytics offering. It should come as no surprise that these are the first API Services to be announced given the focus of Watson on data analytics over the past year. The four new API services include:
- The Natural Language Processing (NLP) API Family: Think of this as Siri or Cortana on steroids where the focus will be on a specific subject area rather than someone asking Watson for the lottery numbers. The press release cites an example of a technician asking Watson about an unusual vibration. Watson will then use its database and other data that it can gather in real-time to analyse the vibration and suggest both a cause and an action to solve it. Another example is the recent launch of The North Face retail site using NLP with Fluid XPS and Watson.
- The Machine Learning Watson API Family: This starts with a set of data around a subject and uses new data to continually reassess what it knows and provide suggestions as to what actions need to be carried out. In effect this will set a new bar for predictive analytics.
- The Video and Image Analytics API Family: Industrial safety systems already use imaging systems to monitor safety. This API would enable greater analysis of images enabling otherwise undetected correlations. The press release cite a forklift being spotted in a restricted area. A few days later performance of equipment in that area begins to drop off. Watson might then infer the potential for an accident between the forklift and the equipment. While a good example it is in the area of law enforcement and anti-terrorism where current systems require massive manpower and eyeballs where this could begin to be deployed quickly.
- The Text Analytics API Family: This will enable a range of new solutions that take unstructured data, correlate it and discover new data. The example in the press release describes using calls, emails and tweets about braking issues where Watson determines there could be an issue with a particular model/make. One of the failures of analytics on large volumes of unstructured data has been the inability to find this type of information out using multiple data streams. There are many ways that this could be useful for consumer companies in particular.
These are just the first wave of API Services that IBM is releasing with a lot more due next year as we get into the IBM conference season.
Conclusion
IBM investing in Europe is good news although the actual size of the investment is not given in the press release. Combining Watson and IoT to create cognitive IoT is more about the back-end systems than intelligence in the IoT devices to begin with.
It will be interesting to see how quickly we start to see IBM promoting new products and solutions as part of this investment.
