Make Turing Proud: Next Best Action in Multichannel Conversations

Big Data Evangelist, IBM

In geek culture, this past Saturday was something special. It was the 100th birthday of the late Alan Turing. If you're familiar with the history of computer science, you'll know that Turing developed the fundamental concept of a "Turing machine." This was his mathematical proof that some abstract algorithmic machine was possible that could perform any conceivable mathematic or logical operation.

Using this concept, Turing daringly challenged the notion that there's a hard-and-fast distinction between artificial brains and the human mind. He proposed what's now known as a "Turing test." This is simply the test of a machine's ability to engage a human in a natural-seeming conversation in which the machine, if its identity is concealed, can behave in a manner indistinguishable from a human.

Clearly, this is one of the founding dreams of the artificial intelligence community. You'll find Turing's vision behind Arthur C. Clarke's "HAL 9000" in "2001: A Space Odyssey." In a non-fictional modern computing architecture, you'll also find Turing's test realized in IBM Watson. What's powering Watson's "deep question & answer" technology is an architecture that supports an intelligent system of engagement. Such an architecture is able to mimic real human conversation, in which the dialogue spans a broad, open domain of subject matter; uses natural human language; is able to process complex language with a high degree of accuracy, precision, and nuance; and operates with speed-of-thought fluidity.

How does this relate to next best action, which is one of today's hottest new focus areas in intelligent systems? It's a fairly straightforward connection. At its heart, next best action refers to an intelligent infrastructure that optimizes agile engagements across many customer-facing channels, including portal, call center, point of sales, e-mail, and social. Your customers (all of whom are human, I suppose) engage in a never-ending whirligig of conversations with humans and, increasingly, with automated bots, recommendation engines, and other non-human components that, to varying degrees, mimic real-human conversation.

Essentially, the Turing test of multichannel engagement is whether you can continuously orchestrate conversations across all channels so that they appear to embody a single human-feeling corporate-brand persona.

This is not smoke-and-mirrors trickery. To the extent that you can humanize your channels, you're likely to boost customer experience, satisfaction, retention, and loyalty. The next best action of multichannel engagement must always be to humanize the next thing you say to customers across any and all contact points. How will you individualize, personalize, and naturalize every utterance, even those driven by embedded statistical models, business rules, and other algorithmic logic?

Yes, you must make your automated responses feel more like they're coming from real human beings. But it's even more important that you make your human-delivered responses feel less robotic.

One of the weak points in many customer relationship management (CRM) next best action programs is the dynamic auto-scripting that drives responses by human channels, such as call-center agents and sales clerk. To the extent that the humans in your CRM channels start to sound like they're simply reading words on a screen, your company has become, in the customer's heart, just another big, dumb, cold, impersonal monolith.

You can humanize your human communications in several ways. First, you must implement more agile and adaptive natural-language-processing technology, such as Watson, inline to your next best action platform, driving more nuanced response-scripting that reflects what's happening in real time through intensive processing of clickstream, geospatial, analog-audio, and other streams. Also, you must give your human touchpoints more powerful real-time decision support tools so that they can use their judgment to dynamically fit their response to the immediate situation – and to occasionally go "off-script." And you must allow your human touchpoints to converse with customers through any and all social channels that add value, supplemented by non-human social-response bots in a way that is organic to the evolving human back-and-forth.

What I'm referring to here are a range of emerging best practices known as "conversational engagement optimization." These practices should always be central to your CRM next best action strategy. Within the context of an ongoing relationship, the customer should not have to notice, or care, whether specific responses are coming from a flesh-and-blood human or an automated information system.

Make Turing proud.