Artificial Intelligence Done Right

Insights from Cognitive Search

The influence of Google on search is enormous. For many people, the words are synonymous. Thus, when employees need information from within their enterprise, they expect the search functionality to be identical to Google. For many reasons, it usually isn’t. Michael Cizmar, Founder & Managing Director, MC+A, thinks that’s going to change. He sees a paradigm shift in search technologies, one that revolves around insight engine maturity.

Cizmar identifies five levels of maturity. Level 0 is where most companies are. For them, search is not a priority. At level 1, companies realize they have an access problem, with no formal governance program and no KPIs. Level 2, Relevancy Defined, initiates governance, maps the user journey, and begins to identify KPIs. Relevancy in Context is maturity level 3. This employs ML-based relevancy, NLP query support, and an emergent search interface. The highest maturity level, level 4, occurs when there’s an auto-classification of content, predictive search, and True Knowledge Assist. To help you on your insight engine journey, Cizmar suggests a 7-step approach.

Sean Coleman, CTO and Chief Customer Officer, BA Insight, also has concerns about the value of search. Citing Stephen Covey’s landmark book, The 7 Habits of Highly Effective People (Simon & Schuster, anniversary edition, 2013), Coleman concentrates on the second habit, “Begin with the end in mind.” To do this, Coleman shares what he believes are the five goals you should strive to achieve when implementing search.

His first goal is to have searches return fast results. Second, users should find relevant information on the first page, no scrolling down to locate it. Third, users shouldn’t have to click more than three times. Fourth, users shouldn’t actually have to perform a search, since information should proactively flow to them. Finally, 95% of users should succeed in their searches. He also doesn’t particularly care whether search is known as cognitive search, enterprise search, insight engines, or something else—his five goals are still key.

Living in a Digital World

AI technologies have a place in our digital world that’s not just about search or written documents. Call centers face challenges unique to the spoken word. We’ve definitely moved beyond early phone systems and should make the most of newer technologies. Here again, bringing AI technologies to bear can optimize performance. Chris Caile, Principal Product Manager for Nuance, identifies four key principles that he considers “central to a truly great caller experience.” Digital, he notes, has not killed IVR (Interactive Voice Response) systems.

To deliver value, the IVR should enjoy an intuitive, personalized first point of contact. At the very least, Automatic Number Identification (ANI) tells you the identity of your caller. Voice biometrics is another possibility. Once connected, your IVR should deliver natural, conversational interactions. Callers’ expectations are high; they’ve become accustomed to talking to machines such as home assistants and smart speakers using natural language. The third principle is to anticipate a caller’s need. AI excels at this. Finally, connect your IVR and digital channels. Omni-channel customer engagement is becoming the norm.

Cognitive and AI technologies are an integral part of customer engagement, according to Anand Subramaniam, Senior Vice President, WW Marketing, eGain. Using AI to improve customer service involves a system that understands and answers the customer’s concerns. Virtual assistants should use reasoning technologies to guide customers through complex interactions. Preferring AI reasoning to rule-based systems aids in decision making. Technology isn’t everything, however. Complying with the regulatory environment is also crucial to success.

Do It Right

To do AI right—and possibly to appease the CEO demanding AI right now—consider the myriad of responses to implementing AI technologies in a meaningful, sustainable fashion. Be clear on the ultimate goal and the immediate problem crying out for a solution. Consider which of the AI technologies, or which combination of them, provide the best answer to your individual situation. Then you’ll know you’ve done AI right.