Watson—IBM’s cognitive computing service that plays Jeopardy, assists doctors in diagnosing patients, and helps financial service professionals recommend products to clients—now has a new role. Watson will help IBM
clients shuffle through reams of data expected to be generated by the masses of connected objects coming online as part of the Internet of things. Watson’s skills will be made available as part of the services offered to customers who come in through IBM’s IoT Foundation cloud product.
Those services will include an API that lets computers understand people when they talk, machine learning and predictive analytics services, a video and image recognition service, and a textual analytics service that can associate meaning from unstructured text-based data. The service will enable clients to build products with completely new capabilities and user interfaces.
For example, a mechanic might be able to ask a computer what step comes next as part of a complicated engine repair or another client might be able to implement predictive maintenance on a manufacturing line, recognizing when machinery is about to break down.
Big Blue is also creating a global headquarters for the Watson Internet of Things division, based in Munich, where 1,000 IBMers will be based. The Munich headquarters will also be one of eight new “experience centers” where IBM
clients can go and see firsthand how to use Watson’s services to build their products.
Like many IBM announcements, there’s a lot of fanfare for what feels like somewhat marginal news. After all, back in March IBM announced it was investing $3 billion in creating an Internet of things division, and in September it named Harriet Green, formerly of British travel firm Thomas Cook, as the head of the IBM Internet of things business unit.
And given that IBM says that there are more than 9 billion connected devices operating today, and that those devices will generate 2.5 quintillion bytes of new data every day over the next 15 years, it’s hardly a surprise that IBM would involve Watson and its artificial intelligence capabilities to parse that data. So what should we focus on with IBM’s Internet of things strategy?
For that, Fortune interviewed Green on her 64th day on the job to understand what mattered and what was more IBM press-mongering. In hiring Green, who is based in London, and setting up the Watson headquarters in Europe, IBM is deliberately placing its assets where much of the actual action is happening around the industrial Internet of things. Governments in Europe so far have been far more interested in investing in projects for smarter cities, while factories and industrial clients in Europe have been quick to investigate new connected technologies in what many call Industry 4.0.
Green tends to go straight to the practical issues. She says that 90% of the client’s data is unusable for traditional computers because it consists of elements such as doctors notes on medical files or social media entries that can’t be classified in a useful format. But Watson can make use of that data—sometimes in startling ways. For example, before Green had arrived on her first day, her future employees at IBM had run her publicly available speeches, social media posts and anything else they could find through Watson to figure out what kind of boss she would be.
“At first I thought, this is kind of scary. But then I realized, my God, this is me,” she said of her future employees’ sleuthing.
That puts a lot of potential power at IBM’s fingertips, which Green was careful to stress that IBM doesn’t treat lightly. Green touted IBM’s security and the respect with which it treats its clients’ data. It actually helped initiate one of the first clouds for the Internet of things that acted as sort of data escrow. The cloud product acted as a neutral, third place for companies that were working together to securely share data without having to grant each other access to the other’s data center.
IBM has 48 data centers around the world and the software capabilities to ensure that data is stored in a specific geographical location if a country’s regulations require it. This is becoming more and more of a concern for the tech industry thanks to data privacy laws.
But the most interesting bit of Green’s discussion wasn’t about security, or the 750 patents that IBM has that are associated with the Internet of things. It was about IBM’s strategy. Green sees IBM pursuing a vertical, industry-specific approach to helping clients adapt to the Internet of things. Thus, IBM has people working with clients in automotive, retail, insurance, those helping on smarter cities, etc.
This runs counter to many of the broad approaches taken by startups trying to break into the industry with a new platform for the Internet of things. Companies such as Samsara, Helium, Ayla and others are starting with a horizontal approach aimed at connecting sensors in any industry, while many of the existing companies, like GE or Cisco, are trying to come at the market with products defined for a specific industry. In many ways it’s the battle between tech and domain expertise, and since the Internet of things requires both it’s worth seeing how this plays out.
Perhaps IBM will end up making a horizontal technology buy if it sees something it desperately needs. Green won’t rule it out. “We are looking assertively at acquisitions,” she said.
For more on the Internet of things, watch this Fortune video:
Make sure to subscribe to Data Sheet, Fortune’s daily newsletter on the business of technology.