On January 9, 2014 IBM announced that it was forming a new organization (the Watson Group) and investing a billion dollars ($1B!) in future Watson development. The company also announced three new Watson cloud services—The Watson Discovery Advisor, Watson Analytics and IBM Watson Explorer—as well as the availability of something called IBM Watson Foundations.
In its press release, IBM described Watson Foundations as “a comprehensive, integrated set of big data and analytics capabilities that enable clients to find and capitalize on actionable insights.” The company went on to say “Watson Foundations includes business analytics with predictive and decision management capabilities, information management with in-memory and stream computing, enterprise content management, as well as information integration and governance. Packaged into modular offerings, organizations of any size can address immediate needs for decision support, gain sustainable value from initial investments, and grow from there.” And I have to be honest, after reading this several times, my reaction was: “Huh?”
To delve deeper into what Watson Foundations is, I was put in touch with John Hagerty, a former technology research analyst at Gartner who is now IBM’s program director for Big Data & Analytics. John provided a very clear definition of what IBM Watson Foundations is and what it means to IBM customers.
What is Watson Foundations?
According to Hagerty, the answer to the question requires an understanding of what IBM’s Watson computing environment is; it is also necessary to understand certain analytics use cases and related technology zones. “Succinctly,” according to Hagerty, “IBM’s ‘Watson’ is a cognitive supercomputing environment that uses natural language input to derive answers to questions using unstructured data.” As for Watson Foundations, “this is a combination of use cases and underlying database, data management [and] analytics technologies [zones] that are used for inquiry and analysis of information that can be found in structured as well as unstructured databases.”
Hagerty’s description of Watson was pretty straightforward, but I had no idea what he was talking about when he described Watson Foundations use cases and zones. So he went on to explain these concepts using types of queries as examples. According to Hagerty, Watson Foundations helps IBM customers perform certain types of analytics. For instance, a customer may seek the answer to the question “what is happening?” To get an answer to this question, a discovery and exploration activity must take place. To do discovery and exploration, data needs to be prepared and placed into an “exploration, landing and archive zone.” This particular zone is comprised of information management, governance and other technologies needed to structure and manage this zone. The combination of a use case with an underlying zone is what comprises Watson Foundations.
Another use case and zone example could be “why did something happen?” This type of query needs database and zone support that includes information ingestion and operational information zone as well as an enterprise warehouse datamart and analytic appliances zone. So the combination of this use case with underlying zone support comprises another Watson Foundations offering.
Yet another use case and zone example, Hagerty continued, is “what could happen?” To answer this type of question an enterprise may need to analyze data in real time, requiring an underlying zone that could support real time analytics—this zone may even use an in-memory database to get answers more quickly. The combination of this use case and the underlying supporting technologies needed to build a real time analytics environment makes up another Watson Foundations zone.
Okay—So how are Watson Foundations related to the Watson Cognitive Computing Environment?
At this juncture, I now better understood how Watson was positioned, and I better understood the Watson Foundations concept, but I did not understand the relationship between the two. So I pressed Hagerty for a clearer picture of the connection between Watson and Watson Foundations. Ho responded to how Watson and Watson Foundations are related by describing how they will most likely merge over time. “Watson is a cognitive computing environment that, at present, works on unstructured data” said Hagerty. “Meanwhile, Watson Foundations performs descriptive, diagnostic, prescriptive and predictive activities on structured as well as unstructured data,” Hagerty explained that, "today, Watson is a question and answer environment, while Watson Foundations is an inquire and analyze environment.” He went on to forecast that, over time, the two environments may blend together such that Watson cognitive computing environments would continue to do cognitive question and answer activities, and Watson’s cognitive fabric could overlay traditional analytical queries.
So, the answer to how Watson and Watson Foundations are interrelated resides in this cognitive fabric concept. By integrating cognitive functions with traditional query and analytics environments (and potentially with the Watson natural language interface), the new types of use cases will be created over time. And, with a natural language interface, human-to-machine interaction could be simplified and improved.
At first, mixing IBM’s fast growing big data and analytics marketing efforts with the Watson cognitive computing environment made no sense to me. But now I look at both environments and I see the commonalities. Watson and Watson Foundations are both environments that can cull through very large databases, enabling customers to make faster, better, more informed decisions. Each uses different approaches to get at different types of answers, but the very bottom line is that both serve a common goal: to help customers make more informed decisions.
As a final note, during the course of my interview with John Hagerty we discussed IBM’s approach to marketing Watson Foundation solutions. I know from many years of following IBM’s efforts in analytics that IBM knows how to build analytics systems; it has a vast portfolio of analytics software, infrastructure, database/data management tools and utilities and systems that can be specially tuned to process specific analytics workloads. But this is not the discussion that IBM wants to have with its customers—instead, according to Hagerty, IBM wants its customers to focus on what information they need to make more informed decisions more quickly. The underlying technologies are enablers, but they should not be the focus of the discussion with customers. Instead, IBM wants to focus its customers and prospects on the types of solutions they need rather than on how to build environments to deliver those solutions.
As I see it, with Watson Foundations, IBM is positioning to usher in new approaches to analytics that will blend traditional and cognitive approaches to formulating and conducting queries—I see no other company in the market marching down this path. If you are aware of competitors who are blending cognitive and traditional approaches to analytics, please comment on this article and add your voice to the conversation.