IBM CEO Rometty: Watson, cognitive computing mainstreaming begins

IBM CEO Rometty: Watson, cognitive computing mainstreaming begins

Summary: IBM's Watson Business Group has launched and the plan is to take cognitive computing mainstream. How much patience will IBM get as Watson exits startup mode and goes commercial?


Positioning the launch of the IBM's Watson Business Group as the beginning of the mainstream cognitive computing era, CEO Ginny Rometty said Big Blue's technology will transform customers' industries, institutions and society.

rometty mug
IBM CEO Ginny Rometty

Rometty's scene setting was at least partly aimed at skeptics and the commercialization challenges with Watson so far. Rometty noted that Watson "is at the very beginning of its journey."

"We don't create new units very often but when we do it's because we see a big shift ahead," said Rometty. "This is a new era of machine/human collaboration."

Rometty said there have only been two computing eras. The first was tabulated---computing was initially designed to count. Then the programmable era arrived with the mainframe and everything that followed. "Everything that is out there today is programmed," said Rometty, referring to PCs, tablets and smartphones.

The third era is cognitive and systems that learn. She said 2011 and Watson's Jeopardy debut "brought a new species" of computing. "By design it gets smarter over time," said Rometty. "It can find a needle in the haystack and knows the haystack. It understands the implications of your questions."

Rometty said that the challenge is taking Watson to a new level with customers and clients. Instead of noting the challenges adapting Watson to healthcare, she said that oncologists at leading research hospitals have dedicated two years to working with Watson. "The greatest testament to Watson is that they are dedicated to it," she said.

Also: IBM forms Watson Business Group: Will commercialization follow? | Dealing with the data deluge, and putting the 'information' back into CIO | IBM Watson: How the Jeopardy-winning supercomputer was born, and what it wants to do next | DBS Bank is IBM's first Watson customer in Asia

Mainstreaming Watson

Michael Rhodin, the new leader of the Watson Business Group, said "mainstream is where we're headed." The Watson team, which will now be 2,000 strong, was in startup mode for the last two to three years.

Watson Business Group chief Michael Rhodin

"IBM will launch new products and new capabilities that will prove what we mean by cognitive computing," said Rhodin. "What you've seen so far from Watson is just the tip of the iceberg. Watson is just getting started. It's learning."

To mainstream Watson, IBM is trying to create an ecosystem of businesses built on Watson and its application programming interfaces. "We want partners," said Rometty.

IBM's Watson division will be focused on the following:

  • Transformational efforts: How can cognitive computing revamp industries? In healthcare, Watson can recommend treatments, integrate information and help teach medical students. Watson can serve as a partner to physicians.
  • Enterprise products: IBM plans to make Watson more scalable with fast deployments to deliver relevant answers to customers.
  • Ecosystem: Building businesses and recruiting developers to work with Watson's technology.
  • Foundational technology: Watson technology will be used for IBM's analytics portfolio.

At the kickoff, IBM launched two cloud services that revolve around Watson, but many more will come.

The challenges

IBM is pushing a new way of computing; in the enterprise and with services, there will be implementation potholes. In that context, it's not terribly surprising that IBM hasn't landed a ton of revenue from Watson.

In other words, the last two to three years have been a startup period and experimental phase for Watson. However, once Watson is implemented in healthcare institutions, it can scale quickly.

Institutions like Sloan Kettering are essentially development partners for IBM and Watson. The exchange of intellectual capital isn't easily monetized. However, once Watson is rolled out to healthcare systems broadly in a more turnkey way, IBM's revenue growth can follow.

As IBM expands Watson's use to other industries, the same implementation trial and effort phases will happen. Scaling Watson across multiple verticals won't be easy or quick.

That reality will also present another challenge for IBM: investor patience. There will be a natural impatience with the Watson unit's growth and revenue contribution. Why? IBM's hardware business will likely struggle to grow in the future due to cloud computing. On the software side, software as a service means a licensing shift for IBM. And for services, the cloud may just mean that you need fewer IBM consultants to implement technologies.

The Watson Business Group will be expected to offset those hurdles and deliver quarterly. If the Watson Business Group delivers $1 billion in 2018 or slips a year to 2019, it will still be successful. Wall Street won't be as charitable if IBM can't hit its numbers.

Topics: Big Data, Cloud, Enterprise Software, IBM

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • the unfortunate reality of the female CEO

    The Yahoo CEO has an unfair advantage over IBMs CEO and the companies fortunes will follow.
    • I think you are being unfair (to be polite)

      to both of them.

      The companies they are in charge of will perform, or not perform, based on their decisions.

      Not how you think.
  • Maybe

    but then there's always plastic surgery. /s

    May I suggest that action and programming will become one function, initiated and controlled by voice.
    Take health care.
    " Why was Idiopathic Pulmonary Fibrosis not included in Diagnostic Ranking?"
    "No Genetic framework was present. No confirming antigens were present. "
    "Expand diagnostic rankings to include non-pulmonary potentials."
    "No other pulmonary inclusions are possible at this time. Additional rankings now include Boete Fungus. A Upper Right Quadrant biopsy is needed for confirmation. No other diagnostic test can confirm."
    "Suggest treatment protocol for Primary Diagnostic Ranking."

    And so on.
    The point is that Cognitive Computing will become very possible in the fullest sense in short order-within 5-7 years, with very rapid improvement following, to the point of 99%+ function success. Non-linear extrapolations WILL be achieved after a "learning curve" based on the empirical performance of millions, billions, trillions (and up) of "experiences" by Users.
    The future of Cognitive Interactive will not just be answering questions by searching for facts, it will be finding answers to questions we haven't thought of yet, AND learning how to ask the questions.
    Having enough sense to utilize the answers is another question.
    "Recommendations, Watson?"
    • Straight out of marketing department

      Software engineer would just look at this and think WTF.
  • Watson was no breakthrough

    Personally, I think Watson was a lot more about marketing by IBM, and a lot less about any kind of breakthrough in AI. The big ballyhoo about Watson is that it won Jeopardy. Well, if you look closely at how that game was played you come away less than impressed.

    For all of its processing prowess, Watson has no speech recognition capabilities (amazing - I know), so this is how they handled the questions. As soon as Alex Trebek began reading the question, the same question was sent electronically to Watson. In relative terms, Watson had received the question before Mr. Trebek finished pronouncing the first consonant of the first word of each question. So as far as thinking/processing times are concerned, by time the other contestants could hear what the question was, Watson had had the equivalent of around 10 minutes to ponder it.

    With this kind of advantage, you certainly hope Watson would have won the game!
    • Actually, that would be about 3 seconds. Not 10 minutes.

      And I'm not sure about even that. The test was setup as fair as they could get it.
  • The biggest problem for Watson is that most people think they are smart!

    It takes above average inteligence to even notice that other people or machines are smarter than you. Stupid people always think they are smarter than everyone else.

    The human brain does not have enough computing horsepower to be a Doctor.
    That is all we have had available in the past so people are used to the idea.

    The only way we are ever going to approach "Star Trek" medicine is with computers like Watson. Doctors will be just medical technicians working for the computer.