How to cross the conversational commerce chasm

There are two sides of the chasm when it comes to conversational commerce.

One side is the long-standing group of professionals who have invested hundreds of billions of dollars and countless man-years developing and implementing enterprise-grade conversational commerce platforms. The other group (which may not even think it’s in the customer care business) consists of the rapidly growing ranks of service developers heeding Mark Zuckerberg’s siren song to go forth and multiply the number of bots on messaging platforms by the number of available outlets.

The customer care professionals have a long history of rapidly recognizing intent and responding to it in a way that is both pleasing to the customer and cost-effective for the company. For more than a decade, that has meant finding and striking the right balance between automated self-service and human interaction. Their efforts have brought natural language understanding (NLU) to the intelligent assistance platforms that power speech-enabled voice response units, virtual chat agents, and text-based prompts that appear in “screen-pops” on the screens of live customer service reps.

The relative newcomers cut their teeth on RESTful APIs, agile development, scrums, and hackathons. They measure success by how quickly they can move from identifying a need for a new bot and its introduction as a friend on Facebook Messenger, WeChat, or Kik; a new skill for Amazon’s Alexa; or a virtual team member on Slack, Spark, or Kik.

The fruits of their labor are “simple” (which I’ve put in quotation marks for a reason) messaging bots that often take a one-and-done approach to task completion. They order a pizza, schedule a meeting, or deliver a news report, then move on. Conversation is not necessary.

The former has already made big investments in natural language understanding platforms that enable both customers and prospects to indicate intent using their own words. They have also identified, indexed, and aggregated multiple feeds of data and metadata that must be brought to bear in commercial conversations. The latter benefit from open source platforms for building purpose-driven knowledge graphs from scratch thanks to free access to the universe of machine-readable data.

Today, the two communities blithely coexist. They live in parallel universes. The bot makers think the customer care professionals are making things too complicated, while those customer care pros see the botsters as an existential threat, luring customers from their captive CRM systems into the freer-feeling (though equally captive) world of social media. The schism is a pity, because the two communities can learn from one another and leverage their respective strengths. Here’s what they can do to cross the chasm.

1. The opposite of a simple chatbot is not a complex one

Rather than conversational interactions, a high percentage of messaging bots limit the number of responses or inputs by making them highly structured and directed. They feature multiple choice questions, radio buttons, or other form of directed dialogue. Customer care professionals regard this as a giant step back to the days of crappy interactive voice response (IVR) systems that let callers “Press 1 for sales. Press 2 for service.”

But the opposite of a simple bot does not have to be complex. It is more fruitful to think of messaging bots as an express lane into the a service platform that leverages prior investment in NLU, along with machine learning (ML), knowledge management (KM), and, ultimately artificial intelligence (A.I.). The presentation layer may feature a single icon or even an emoji to launch a task or take control, but its core role is to mask the complexity of the sort of pattern matching or deep analytics that is taking place under its hood.

Thus the role of a “simple” messaging bot is to leverage investment that has already been made in conversational self-service and intelligent assistance platforms.

2. Message bots become highly personalized advisors

In the Intelligent Assistant Landscape most recently issued by Opus Research, in April, the bottom row divided the world into four categories.

• Mobile & Personal Assistants, like Siri, Alexa, GoogleNow, and others, help people take control of their personal devices and services.
• Personal Advisors, like many of the messaging bots, as well as health coaches or financial planning tools, build specific domain expertise and provide advice or motivators to help people reach their goals.
• Virtual Agents and Customer Assistants are the avatars, chat agents, or other virtual agents that provide conversational self-service for businesses and their brands.
• Employee Assistants and Advisors are tools to support intra-company productivity, efficiency, and collaboration and will support a community of employees.

In short order, the vast majority of chatbots will fall into the “personal advisors” category. Developers will strive to make them available to the broadest number of individuals through the most popular messaging platforms. They will do so by making them into trusted providers of highly useful services. They will be permanently resident on their users’ friends list or directory and, while they may not be conversational, they will be easy to engage, context-aware, and armed with the most recent and most accurate answer to search queries and questions.

In other words, they are leveraging the best of the enterprise virtual agents and customer assistants category. That has tremendously positive implications for “classic” intelligent assistants because they have incentive to stay current as it becomes clear that they are engineered to be a single source of correct answers and recommendations.

3. Brands are crossing the chasm

Individuals expect conversational commerce to be free. It costs nothing for them to have Siri on their iPhone or GoogleNow searching for local businesses. It took almost no time for Amazon Echo owners to discover how to order household goods, clothing, or entertainment through Alexa. Their choices threaten to disintermediate the major brands — including credit card issuers, banks, airlines, and major retailers. Large companies in these categories have already made heavy investments in conversational commerce and intelligent assistance. They are defining brands that will cross that chasm.

The new term will be ‘optichannel’

I first heard the term “optichannel” at a USAA briefing, and it resonated with me immediately. Beyond the play on the old term “omnichannel,” it appears to be the elision of the phrase “optimal channel.” That literally means the channel that is best for the customer’s purpose at the time. But the word “optimal” also starts with “opt,” which means to choose. Thus it simultaneously invokes the idea of a “channel of choice.”

“Optichannel” establishes a higher bar for customer care and user experience. Combining options with optimal puts customers at the center of the customer care equation again, something the CRM types and the bot developers can agree on.