IBM is going through a painful transition.
Just this week, the 104-year-old company announced a miss on its quarterly earnings — the fourteenth quarter in a row that IBM posted lower revenues than a year ago. Wall Street analysts remain wary, and IBM stock is getting pounded.
But IBM insists there’s a bright spot: Its cloud computing business, which it says grew 65% in this last quarter versus the same period in 2014.
At the same time, companies like Amazon, Microsoft, and Google have been aggressively growing their own cloud businesses, slowly but surely eating the market. Amazon Web Services alone is a $US7 billion beast of an IT business.
But IBM Corporate VP of Cloud Robert LeBlanc thinks that the company has a secret weapon in the cloud wars: Watson, the company’s hyper-smart, Jeopardy-playing supercomputer.
A secret weapon
Basically, LeBlanc says, the whole “cloud” thing has moved on from just providing IT infrastructure. IBM has that part down, in a variety of ways, across a variety of situations. So do Amazon, and Microsoft, and Google.
IBM is also making a big bet on “hybrid” cloud, which is where it lets customers mix-and-match their own data center infrastructure with its cloud services. But Microsoft (in particular) is also in that space.
That’s why Watson is such a big deal, LeBlanc says, and why it gives IBM a “leading edge” in the cloud. IBM calls it “cognitive computing.”
By letting developers use Watson, IBM can offer a place for companies to build smarter apps that use Watson’s famous capabilities of understanding and processing natural human speech and providing an answer.
When IBM Watson was on Jeopardy, it had 5 “key modules,” LeBlanc says. Which is to say, it could process 5 different types of data.
Nowadays, Watson has 28 key modules, meaning that it can understand, interpret, and answer questions in really logical and useful ways, LeBlanc says. And all of those modules are open to developers.
“We’re making it easier to get access,” LeBlanc says.
And it’s helping developers tackle a serious problem.
In computer science, there are two kinds of data: structured data, which is the kind of easily-organised data you might find in your ordinary Excel spreadsheet. And then there’s unstructured data, which is stuff that’s hard to organise, including video, audio, the contents of emails, and so on.
Unstructured data can be hard for computers to make sense of. A human may recognise that a three word email, like “Did it work?” has to do with a particular project. But a computer lacks the context to figure out what that could possibly mean. It can tell you how many letters were in each word, but it can’t tell you why you sent them.
Over 85% of all data that IBM deals with is unstructured, LeBlanc says. It’s such a problem that its developers have taken to calling it “Dark Data.”
Watson is getting better at guessing the significance of our emails, phone calls, questions, and videos. Since it’s hosted in the IBM Cloud, companies don’t need to have their own supercomputer on premises to take advantage.
Microsoft has started working on this problem with the Cortana Analytics Suite, which takes a similar approach to helping customers sift through their data likes sales and revenue forecasts. But Microsoft hasn’t yet let developers in on the action by letting them integrate Cortana in the same way.
But IBM’s been doing AI for a long time. Internally, IBM is using Watson to build its own solutions for healthcare, providing doctors with insights into how effective their treatments may be. It’s even putting a Watson-powered doctor consultation service in CVS stores.
So while Watson may not be great at guessing your personality just yet, IBM is betting that the ability to simply and cleanly get answers from a piece of software by using natural language is the next great frontier in the cloud — a frontier that the competition has only begun to explore.
“We really believe it’s the next generation,” LeBlanc says.