Based on an informal assessment of my email inbox and the marketware I receive in the course of Web browsing and app use, “chatbots”, “answer bots” or just plain “bots” has become the accepted term for what Opus Research has long called “Intelligent Assistants” or “IAs.”  Treating “IA” as a thing was an important step for us as we attempted to bring a coherent, single concept to capture the automated entities that span the tens of thousands of bots on messaging platforms, mobile virtual assistants (like Siri, Alexa or Google Assistant), branded virtual agents like the Army’s SGT STAR, JetStar’s “Ask Jess”, or Tangerine Bank’s Inge to name but a few. As I put it in a number of public presentations: “IA is our consumption model for AI” – meaning that we humans get the most benefit from artificial intelligence (AI) when it takes the form of intelligence augmentation (IA), better known as “intelligent assistance.”

One of Opus Research’s mantra’s is “Words matter!”. However, in this instance, our attitude moved from mild disgust and denial (because of the negative connotations of “spambots” and other forms of malware deployed as bots) to acceptance because of near-universal understanding of the meaning of the term. Even though Dictionary.com’s first definition is “the larva of a botfly – an internal parasite of animals [especially horses],” it is more likely to be recognized as a shortening of the word “robot,” and recognized as a computer program or “microservice” embedded in a Web site or mobile app, that can talk or chat like a human. Their proliferation, mass deployment and, most importantly, general acceptance is driving a wave of re-thinking, re-architecting and ultimately re-design of the IT infrastructure, data management processes and organizational structures that support both customer care and digital marketing.

Customers Are Already Digital; Brands Are Catching Up

Every individual with a smartphone and ID on a social or messaging platform like Facebook, WeChat, WhatsApp, Line, Slack, or others has gone digital and – at least when messaging and text-based chat is involved – “conversational”. Meanwhile, brands, large and small, who are already vested in omnichannel contact centers and IVR systems must treat each new conversational platform – starting with mobile apps – as a completely new channel, replicating common functions (like “Where’s my package?” or “What’s my balance?”) and re-rendering it appropriately. Lately, that has meant building an Alexa skill, Google Action or building a “bot” for Facebook or Apple.

The success of “bots” has given rise to existential questions like, “What does it mean to be ‘omnichannel’?” “Are we witnessing the death of the Contact Center?”, “Are bots replacing humans?” and “Who’s got the budget for digital transformation?”

Because the ultimate objective is for every business or brand to optimize user experience, the above questions are interrelated, interdependent and span all business units. Yet, like so many aspects of digital commerce, it starts with the underlying data (Big Data and Analytics if you will). To understand this, it is important to do some level setting. The aspects of AI that have the most impact on creating a pleasant and effective user experience are:

• Speech and text analytics (the key enabling technology that makes categorization and identification of intents possible)
• Natural language understanding (NLU), and
• Machine learning (ML)

Collectively this troika make it possible for a brand to detect patterns that support prediction of the purpose of each contact or “moment”. NLU resources assign each call to a category that enables a company to understand and respond to an individual’s intent quickly. ML – which is often regarded as a superset of both analytics and NLU – plays a dual role. As “learning” implies, it is the mechanism for constant refinement and improvement of results. Yet the preponderance of implementations we are aware of should be considered “supervised ML”, the tools and platforms that are in use to support a brand’s ability to learn from past activity make humans the ultimate arbiters of correct responses. They are the guard rails against an automated system “going rogue” like Microsoft’s Kai.

Beyond ROI: The Value of Data for Conversational AI

The overall impact of deploying AI in for customer care and e-commerce is to save money by replacing jobs. Replacing labor costs with automated handling provides the basis for ROI calculations and business plans. Real world experience shows judicious deployment of conversational AI makes it possible to save money while, at the same time, improving customer experience, routinely boosting CSAT and Net Promoter Scores (NPS) because customers can avoid waits, take control of their experiences and get results quickly.

In spite of promising a long-term decline in contact center employment, conversational AI does not spell “the death of the Contact Center.” As we will document in other pieces, voice-based conversations with agents or advisors are the basis of decisive moments between brands and their customers (not to mention prospects). In the mean time, we are witnessing nothing short of the redefinition of work as we know it. Not to be lost in discussions of “The Future of Work” is the fact that bot deployment creates great opportunities for brands to handle a massive number of short conversations. Existing Contact Center metrics and infrastructure had discouraged such bursty, asynchronous “sessions”; but now they can be actively encouraged, to the benefit of brand and customer alike.

The short term impact of conversational AI, or IA, is to give employees two very important tasks. As subject matter experts (SME’s), the “best” customer service representatives, sales people or technical support engineers provide the initial training material for Intelligent Assistants or bots. Tools and platforms for developing and deploying bots are designed to ingest input in the form of transcripts from chat sessions or spoken conversations. Then, as the system matures, the same subject matter experts or their counterparts have a long-term role to play as “supervisors” or arbiters of truth to make sure that the designated “right answers” as determined by the IA platform, are the best that they can be.

The future of digital marketing, sales and customer care is clear, and conversational AI in the form or IA (okay “bots”) is a big part of it. Discussion of whether bots and AI spell the end of contact centers, mobile apps, e-commerce Web sites and omnichannel support strategies, in general, amount to strategic decisions for customer care professionals and those in charge of digital transformation and technology innovation. In the best case, even though contact centers are treated as an expense centers today, their databases, analytics systems and key personnel promise to be fundamental building blocks for informing bots and conversational intelligent assistants for the foreseeable future.

Join us at the Conversational Commerce Conference in San Francisco (Sept 12-3) for executive perspectives, panel discussions and case studies addressing the challenges of implementing conversational AI to support customer care, marketing and digital transformation. You’ll also find great insights by joining Mitch Lieberman, Abinash Tripathy (CEO and Founder of HelpShift) and me for the Webcast “Speaking Your Customer’s Language: How to Build a True Conversational Messaging Platform.”

Categories: Intelligent Assistants, Webcasts, Articles