Data Mining

IBM's Big Blue hope: How far can Watson go?

IBM Watson is at that tantalising stage of life. Its approach is that of a startup but it’s incubated by the world’s most historic IT brand. It’s in a space where competing companies and approaches overlap and nobody knows who the winners and losers will be. The promise is to answer some of the world’s biggest questions but how satisfying will those answers be? It’s all very promising but who knows what sort of financial numbers Watson will generate. Questions, questions...

What is clear is that IBM is putting a lot of faith in Watson, named after its founder Thomas Watson and backed with $1bn of Big Blue money plus talent and other resources. The mission: no less than to change the world through cognitive computing: software that learns as it goes and can use supercomputing power to answer complex questions and change the world in both small ways and large.

A small way: Watson-generated ‘cognitive cooking’ recipes, analysed by seeking surprising synergies of ingredients and fed to journalists like me at a recent press conference with unusual chemically-inspired combinations such as Belgian bacon pudding. Watson is also being used to answer questions posed by fans at the Wimbledon tennis tournament. On a much higher echelon, it’s a tool for oncologists and, if it’s not yet ready to find a cure for cancer, it’s likely that cognitive computing or a similar approach will play a role in doing so.

I spoke to Steve Gold, CMO of the Watson initiative, to get a catch-up on progress.

“We announced Watson in January 2014 so we’re 18 months into it,” he tells me by phone. “That’s when we made the decision to formalise a business unit, which IBM very rarely does and had not done for 20 years, because we saw the future of cognitive computing.”

Just as IBM minicomputers were centred in Minnesota and the Personal Computer originated in Florida, Watson has its own home outside the Armonk, New York mothership, in Austin, Texas, says Gold, a former SPSS executive. He defines cognitive computing as an approach that dispenses with the usual rules and structure of data analysis and where systems “learn and get progressively smarter based on use and interaction and are capable of understanding us in an intuitive way”.

He uses an old standby term that has gone a little out of usage in recent years to describe the effect of this – a “paradigm shift” – and says that Watson lends itself to a particular class of deductive problem: “[It works] in the same way [as when] you go to the doctor feeling unwell and the doctor checks for symptoms and tries to figure out if it is a cold or a virus.”

Watson is now in use across 17 industries in 36 countries, Gold says, and has gone from having three partners to 300 in that year and a half but it has also evolved in its technical capabilities.

“It cannot just uncover patterns that say ‘If a particular board member should leave, should I sell the stock?’ but ask more open questions like ‘How do I plan my retirement?’

“We can’t do this alone, there are too many places that we can transform a business or industry. We’re moving very fast and at times it’s a little scary; we’ve seen this torrent of incoming activity where it’s a little bit of a tsunami.”

Gold says that the cognitive marketplace is new so “there’s no historical trend line or baseline for compound annual growth rate” but says the opportunity is vast and will be accretive to historical IT spend.

He’s also keen to rebut the notion that Watson is in some ways a rival to the Hadoop open source framework for collating data, describing Hadoop as complementary but an “underpinning”.

“Cognitive is not a very interesting experience if you don’t have the data – that’s the corpus repository to be used to find new patterns.”

Partnering will be critical, he insists and he turns almost poetical in describing the need for “tinkerers, creative types that dream the wild dream” and what they might achieve when their ideas are harnessed to Watson’s people and “and the number-one B2B brand in the world”.

Watson’s first vertical was healthcare but he sees broad adoption across industries and sectors and points to entrepreneurs such as Travelocity founder Terry Jones who now, as chairman of Wayblazer, is focused on putting the travel agent back into the travel experience. “So I can say I want a holiday where kids can hang on the beach and I can play golf and by the way I want to use the travel miles,” Gold says.

Gold is cagey on to what extent IBM’s relationship with Apple could help Watson’s progress but I can imagine a time when a Watson-powered Apple Watch might have a virtual agent capable of handling requests better than Siri. But Gold will only say this:

“When you look at Apple, they’re probably the best when it comes to the end-user experience and understanding the B2C dimension of business modelling. IBM brings enterprise, scale, the brawn to make this happen and the brains from the complementary point of view.”

How long before Watson becomes mainstream?

“My short answer is it’s happening now,” he says, pointing to 50,000 students at Deakin University, Australia having the chance to access a Watson-powered dashboard to aggregate information about their studies and ask questions, or Thailand’s Bumrungrad Hospital’s cancer care programme.

Another passion is Elemental Path’s ‘CogniToys’ such as the interactive WiFi-connected toy dinosaur that children might ask questions of, such as ‘why is the sun yellow?’

“It gives purpose to my life: I know and will tell my grandkids I did something to help advance healthcare and education,” Gold says. “We kind of have our cake and eat it too: we have the personality of a startup and all the nurturing a parent would give to an up and coming child. [For Watson] it’s not ‘when’ it’s now… and it’s only going to get bigger.”



« A new crowdfunding model for Togo


The wooden microchips and screens of the future »
Martin Veitch

Martin Veitch is Contributing Editor for IDG Connect

  • twt
  • twt
  • Mail


Do you think your smartphone is making you a workaholic?