Driving Toward Cognitive Computing Breakthroughs
In the U.S., it’s a rite of passage to get your driver’s license when you’re 16. If you live in most other parts of the world, you’re probably thinking “That young, really?” Yes, really. My father taught me how to drive in the family car, which had an automatic transmission. He was an engineer, so he also trained me to listen to the gears as they shifted. My mother reminisced about when she learned to drive in a car that had a manual transmission. She also talked about listening to the gears.
One summer, I was going to be driving across the country with a friend from university. Her car was a stick shift. I didn’t know how to drive it. My mother called several car rental places until she found one that would rent us a car with a manual transmission for a day. We spent that day working on gear shifting, putting into practice all that gear listening my father had insisted on. It was a wonderful experience, and I’ve preferred manual transmissions ever since. I’ve even driven a farm truck that taught me the meaning of compound low.
I like the feeling of control that you get from “driving stick.” I can choose when to move from one gear to the next rather than be at the mercy of whoever designed the automatic transmission’s decision on optimum timing for gear shifting. Next on the horizon: Autonomous vehicles. I haven’t experienced driverless cars yet, but they have interesting possibilities.
Although I learned my driving skills a bit in reverse from the actual progression of automotive technology, I find parallels between automotive transmissions and search technologies. We started out being in charge of what computers did. Literally, we are issuing commands and the computer followed those commands. Then we moved from manual transmissions to automatic. Computer code took over from command language. Both simplified our lives. We didn’t have to think about when to shift gears. Although I’m a fan of manual transmissions while driving a vehicle,I wouldn’t readily return to the days of command code.
Now it’s the turn of algorithms and cognitive search. To my mind, autonomous cars and cognitive computing have a lot in common. They rely on technology to get to their destination.
Sean Coleman, CTO and Chief Customer Officer, BA Insight, has a comment that nicely sums up, for me, the value of cognitive search. As one of the goals of search systems, he says, “Users Should Not Have to Actually Perform a Search.” I love this concept. Users don’t have to remember arcane search commands; they don’t even have to think about putting words into a generic search box. Instead, the search system figures out what they need and automatically delivers it to them. It’s the ultimate in push rather than pull technology. Users don’t have to pull answers to their questions, the system pushes it to them. Essentially, it knows what they need to know.
Coleman views cognitive search as having five key goals. Besides the zero-search goal, it is about returning fast results, finding relevant information on the first page, not having to click more than three times, and making sure search success is achieved by 95% of users. I’m told that autonomous cars are programmed not to exceed the posted speed limits, but I’m pretty sure that there’s no speed limit for receiving search results. Coleman puts it at between three and six seconds. It takes me longer than that to back out of a parking space.
I would think it’s axiomatic that relevant information should appear immediately, but relevancy is frequently easier to talk about than to deliver. As Coleman points out, it takes intelligent tagging, good metadata creation, and effective text analytics. With enterprise search, in particular, personal izing search results leads to greater relevancy. That’s where BA Insight’s machine learning capabilities come in.
Clicking around randomly sounds like grinding gears to me, so the “not more than three clicks” rule makes a lot of sense. A measurable goal of 95% of users succeeding in their searches is impressive, and Coleman presents some interesting ideas about measuring search success.
Shift Into Reasoning Technologies
Over at eGain, cognitive and artificial intelligence (AI) technologies for customer engagement are tearing up the roadways. Anand Subramaniam, Senior VP of Worldwide Marketing, calls them “reasoning technologies” that learn from past experiences to find solutions for new ones. Think of that autonomous car that finds the optimal route from one place to another.
Making good decisions is central to successful businesses. For customer service, this means moving away from rigid scripting and rule-based systems into a more intelligent understanding of the problems customers face. To provide good solutions that quickly resolve these problems, eGain employs AI technologies that put customers on the proper roadway rather than leading into dead ends.
Flexibility is key to good decisions, and eGain uses AI technologies to help out here as well. Its ability to work with incomplete, ambiguous information yet still learn from the past fuels successful customer engagement. Decision making is not a cut and dried process. Thus, it is ripe for AI involvement.
Whether you prefer driving a manual or an automatic transmission car or want to experiment with self-driving vehicles, you should test drive some of the exciting cognitive computing opportunities. Put the pedal to the metal and embark on the journey. You don’t need a driver’s license to head toward cognitive computing breakthroughs.
Companies and Suppliers Mentioned