At a time when old-school giants like Hewlett-Packard, Dell, and EMC are struggling to cope with the decline of the PC and the rise of cloud computing, IBM is betting big on the future of its cognitive computing group, better known as Watson, and relying more heavily than ever on partnerships with developers to bring its futuristic technologies to market.
The goal: $1 billion in annual revenue.
To hit that target, Big Blue last week unveiled a 2,000-person business unit that will draw on the expertise of consulting pros who bring backgrounds in machine learning, advanced analytics, data science and development, and industry and change management.
IBM’s Watson group, once criticized for ignoring developers, has pivoted. It has now enlisted dozens of software vendors and tens of thousands of developers. Watson will soon open offices in San Francisco’s startup-heavy SoMa neighborhood, and its Watson Developers Cloud offers coders the chance to use dozens of APIs at no charge.
In less than two years, the Watson platform has evolved from one API and a limited set of application-specific, deep Q&A capabilities to about 50 APIs powered by dozens of technologies.
Moving so hard into relatively unproven territory makes a good deal of sense for a company that still derives significant revenue from mainframes and other hardware. Despite aggressive moves to reposition the company, investors are still very skeptical. IBM’s stock has lost about 16 percent of its value in the last year, while the tech-heavy Nasdaq has risen about 13 percent, despite the market’s tough few months.
IBM knows it needs to, ahem, think different.
Developers are now critical to the Watson ecosystem
Watson, of course, made its debut in 2010, when it bested a bevy of "Jeopardy" champions in a widely watched TV smackdown.
At the time, it was a computer the size of a minivan, but IBM choose to move its capabilities to the cloud. Now Watson is a complex fusion of big data, analytics, natural language processing, machine learning, and artificial intelligence aimed at extracting useful information from masses of largely unstructured data.
IBM has been selling Watson as a service into numerous vertical markets, including health care, where it is used to help oncologists sift through huge amounts of data on clinical trials and drugs.
Only a few years ago, IBM took heat for building a proprietary platform that didn’t seem to leave room for developers. As IDC analyst Matt Eastwood argued in early 2014, "To succeed, IBM needs to build a different kind of ecosystem."
Apparently, IBM was listening.
"We expect to partner with developers when they are ready to bring cognitive-related applications to market," says Stephen Gold, vice president of the Watson group. The partnership will include revenue sharing for products that actually launch, and IBM will lend a hand with sales and marketing chores.
To date, developers have launched about 100 new applications, and some 400 software vendors, including more than 77,000 individual developers, are working with Watson. IBM will soon launch a certification program in cognitive technologies, Gold tells me.
Late last month, IBM released new and enhanced Watson cognitive services across four areas: language, vision, speech, and data insights. They're meant to reduce the time required to combine Watson APIs and data sets, as well as to embed Watson APIs in mobile devices, cloud services, and connected systems.
Analytics tools for users
Watson is also putting analytics tools directly into the hands of users. Yesterday, it unveiled new offerings in its Watson Analytics line: data-discovery models called Expert Storybooks, new connectors to a range of data sources, and more secure connections to corporate data.
The Expert Storybooks are essentially data-discovery models producing sophisticated visualizations to help users find relevant facts and discover patterns and relationships to make predictive decisions.
With baseball statistics from AriBall, for example, Watson Analytics offers an Expert Storybook to build predictions of player performance for users to get an edge against their fantasy baseball competitors. An Expert Storybook built with the Weather Co. is designed to help users incorporate weather data into revenue analysis; a Twitter Expert Storybook helps analyze social data to, among other things, measure reputational risk.
Whether Expert Storybooks and the like are examples of cognitive computing or analytics isn’t clear -- and probably doesn’t matter from a business point of view. Both baskets of technologies are rolled up into IBM’s analytics business, reportedly the company’s fastest-growing segment.
IBM's betting on Watson as the future of the company
Although IBM would, of course, be more than happy to sell hardware and other software and services to Watson customers, it won’t insist on doing so, says Gold.
How big is Watson? IBM CEO Ginny Rometti has said it could be a $10 billion business. John Kelly, IBM’s senior vice president in charge of Watson and related businesses, told a Bloomberg reporter that he expects it to be a $1 billion business fairly soon. When I asked Gold the same question, he didn’t want to touch it.
In any case, Rometti has clearly bet the company’s future on Watson. One sign of Watson’s importance to IBM is Kelly’s status. A glance at the company org chart reveals he ranks slightly below Rometti herself.
I’ve heard Kelly speak a number of times, and when he says wants to change the world, it doesn’t have the hollow, self-congratulatory feel of Silicon Valley’s young billionaires. At a gathering of hundreds of researcher scientists and academics in San Francisco this week, Kelly talked about why he wants Watson to focus on the world’s most important problems. "What is the price of not knowing? Of not curing cancer or not discovering alternative energy?" he asked.
Kelly’s vision is inspiring. But IBM has preferred a similar vision for decades in one guise and another, usually from its T.J. Watson Research Center's projects. Whether it translates to a new life for IBM this time remains to be seen.