A Case for Sherlock: Artificial Intelligence?
Sir Arthur Conan Doyle’s 1886 fictional “consulting detective,” Sherlock Holmes, was a great mind renowned for his highly advanced powers of observation and reasoning. He was often assisted by Dr. Watson, who was unfailingly loyal, if noticeably less bright. At the end of each thrilling tale starring the duo, the anxious reader would hear Sherlock announce that he had solved the latest mind-bending riddle, inevitably characterizing the solution to his trusty helper as, “Elementary, my dear Watson!”
Sherlock, no doubt, would love today’s hottest trend Artificial Intelligence—as he would be thrilled by its ability to make plain that which appears impossible or difficult to achieve. For the past few years, for example, we’ve grown to love streaming music providers like Spotify, Pandora, and Tidal, which get smarter about our musical preferences the more we use them. More recently, assistants from Amazon and Google have become the coolest gadgets to have, as they quickly serve up answers in response to simple voice commands. Then there are the daily headlines about self-driving cars, that, no doubt, will soon be coming to a driveway near each of us. Not to mention Nest thermostats, Bluetooth light bulbs, etc., which represent other aspects of our daily lives that are becoming more intelligent.
Innovation: B2C Before B2B
But what about life at work? Is your work life becoming magnificently easier, thanks to an enterprise-grade Siri, for example? For most of us, the boring answer is simply no, that it’s more like Watson than Sherlock. So why is AI making such noticeable strides in the B2C realm, but not yet making comparable strides in the B2B part of our lives?
We’ve seen this movie before. Consider Google. We can’t live without it today, but remember when there was a battle of search engines, when fledging companies were competing to be our engine of choice to help us find the most relevant results on the “information super highway”? Those early web crawlers (whatever happened to Ask Jeeves, anyway?) made navigating and extracting value from the still-fuzzy “World
Wide Web” much easier. But did finding stuff at work become easier thanks to those powerful new web tools? No. The enterprise analog to Google and its peers was the enterprise search platform, with companies like Autonomy, Endeca, FAST, and others leading the way and eventually getting acquired for huge sums by mega vendors like HP, Oracle, Microsoft, IBM, and others. But did productivity at work increase dramatically thanks to those enterprise tools? Again, no.
So why is innovation in the B2C world once again—first with search, now with AI—outstripping the B2B space?
The answer, my dear reader, is quite elementary: security.
Security, My Dear Watson
In the case of web search, finding information is easy in some ways. The information is out there. Much of it is tagged. All of it is indexed, analyzed, and enriched. And everyone is allowed to see everything so security is not an issue. The trick is simply finding what’s most relevant. In the enterprise, though, things are much more complex. There are multiple repositories to search. And not everyone is allowed to see everything. Every repository has its own permissions based on user roles. But as people get promoted, transfer, job-hop, etc., the information within the enterprise index must be re-processed, which is no small or easy task. The best-in-class trick to security within the enterprise is to store the indexed information separately from the permissions and to join them dynamically, once the enterprise user submits a query. This “Active Security” model is far superior to early- and late-binding methods that are commonly used and cause latency issues. Clearly, security is a job for Sherlock, not Watson.
So will AI ever have a positive effect on the B2B space, or will it only really help B2C, as we experienced with search?
Keep Your Eye on Machine Learning
Machine learning—more precisely, machine learning-based relevancy—is basically a system’s ability to infer your intent based not only on your explicit action or query, but also based on your implicit indicators. It’s a continuous process—a virtuous feedback loop—of presenting ever more refined answers to queries, based not only on what words you explicitly submit in a query, but also on who you are, where you are, what device you’re using, and what behaviors you’ve demonstrated previously, in response to other queries. This is heady stuff. Sherlock-smart.
In a word, machine learning is fundamental to both search and to AI. It’s what makes search “cognitive.” And cognitive search is the first step on the path to AI. Machine learning—more broadly, natural language processing—is not only making information easier to find inside the enterprise, it’s also the secret sauce inside those cool consumer tools like Siri, Alexa, Google Assistant, Cortana, etc. Just as you have Alexa on your kitchen counter at home, you now are seeing domain-specific search applications at work that provide continuously optimized search results and proactively present answers and alerts for specific
use cases like knowledge management, expertise location, know your customer, anti-money laundering, eCommunications Surveillance, etc. These new apps are, fundamentally, search apps relying on foundational AI technologies. Because these apps are typically presented to the end user from within existing UIs, they are familiar and easy-to-use and, therefore, quickly adopted.
In some respects, therefore, B2B is—finally!—catching up with B2C. Machine learning is what’s powering this cycle of tech innovation and is what will propel AI to make truly Sherlock-like improvements in our lives, both at work and at home. ?
Attivio is the leading Cognitive Search and Insight Platform company. Our Fortune 500 clients rely on us to drive innovation, operational efficiencies, and improve business outcomes.