Conversations are important and well understood. Artificial Intelligence (AI) is important but only partially understood. Stringing all of the words together does not add clarity, quite the opposite. Conversational Artificial Intelligence, better known as Conversational AI, is a misunderstood and misused term. There are too many interconnected complex components to create a single cohesive narrative. For the purpose of this discussion, a conversation is intended to serve a purpose (this is not about chit-chat). Conversational AI neglects important details, excludes the human factor, and ignores key market players and dominance. Human intelligence is critical to business success but is absent, by definition, from Conversational AI discussions. We need clarity. We need to break the problem down (as best we can), then put it back together. Conclusion: Conversational AI, as a term, is going to get in the way of the progress required for Customer Experience practitioners.
Intelligence Is A Component Of A System. AI Is A Subset Of Intelligence.
A system is the combination of a platform (multiple software applications) along with people and processes. Systems help enhance an organizational capability and scale operations. Intelligent systems help organizations scale faster and move toward operational excellence. Making a system more intelligent is achieved one component at a time. Conversational Systems are focused on scaling the organizational capability towards having more valuable and more meaningful conversations between two parties. Each conversation needs to be well-informed, this is intelligence. Sometimes conversations are with people, sometimes they are not. Sometimes conversations are via voice, sometimes they are not. AI does have a role, maybe more than one. We need to be specific about the purpose of each role.
(Where this may seem a bit complex, it is actually a reductionist view based on what is properly represented as a Complex Sociotechnical System – so, if you would like to really dive in, start with this post from Irving Wladawsky-Berger.)
Intelligence requires turning data into information and insight. Intelligent systems require both smart platforms and smart people. A person becomes smarter through learning and experience. A machine (aka software) becomes smarter the same way. Software that acts on insight (algorithmic pattern matching) and is able to learn what works and what does not work, is considered Artificially Intelligent (because there is no human involved in this feedback loop). Thus, Machine Learning is a software application that is able to get better at its job. The output is known as Artificial Intelligence. When the software is smarter and the people are smarter, the system is smarter; an Intelligent System.
The Amazon Affect
Conversational Systems need to support the creation and enhancing of relationships. CRM: Marketing, Sales, and Service is the Enterprise System and discipline charged with the building and maintenance of business relationships. The link is clear. But this is within the context of Enterprise Software. In the consumer Conversational domain, relationships are not part of the discussion (not yet anyway). That is just fine with the folks at Amazon, “right, Alexa?” As noted above, there is more to this discussion than platforms and software. But that is not what Amazon would have you believe it is all about, an assistant, supported by a platform without people; Alexa. Amazon has a definition:
“Conversational AI systems are computers that people can interact with simply by having a conversation, our most natural form of interaction. In short, it is what allows us to talk to voice-driven technologies…”
Amazon owns the branding (thus, the SEO), this does not help matters. When someone searches online for “Conversational AI” the answer (above) may not be what is expected. As a very interesting and important aside, the most sophisticated component on all voice-based systems is the microphone (hardware, no AI). The microphone array will make or break the usability of systems using voice as an input. Just one example of interconnected complexity. When performing a search and looking for answers from businesses and Enterprise technology providers, the definition is a bit different (and might take some scrolling).
“Conversational AI is the set of technologies used by messaging apps, voice assistants and chatbots (non-human) to automate communications between a person (customer) in a scalable and personalized manner.”
The issue, as noted above, Amazon owns the space and they have a big budget. There are two areas of concern. The first is that conversations are not only about voice (verbal communications). Amazon with its influence in the market has made Conversational AI a voice only, or at least voice first discussion. The second issue is that conversations are much more than the words, intents, and emotions. Context is critical. Creating context requires going outside of the conversational system (with AI or not) to build context. Amazon will eventually solve the context issue, but I do not believe it is the top priority.
The Focus Needs To Be On Outcomes And Experiences
Practitioners focus on outcomes and experiences. Vendors and advisors need to help them to keep this focus, making customer (and conversational) experiences better. Conversations are multi-modal. A person may start with voice and then switch to text or video. The response might be a different together, like a message to my phone. It might include AI, but again, it might not. Enterprise Vendors need to exercise caution before jumping on the Conversational AI bandwagon (even if they were there first). A Google search on “conversational AI” will see results that are almost in contradiction. When practitioners are making purchasing decisions, Google is the starting point.
The Amazon definition is not wrong. In the land of consumer voice assistants, it works. The Enterprise needs something different, a systems lead, outcome driven and experienced focused approach. After all, everyone wants a consumer-like experience. The most mature part of AI within Conversational Systems are the recognition of the input (ASR, NLP, and NLU). The future focus needs to be about emotions and intent that can be added to the conversational flow. Amazon will get there, lead the way even, but not in the way the Enterprise needs. There is a lot here, and a lot to be impressed with, offered by many technology companies and vendors. Dear Vendors – highlight your greatness, just not using “Conversational AI” as the baseline. Among all of the AI parts, the standard technology sophistication and rules are also overlooked – Conversational Systems can be, and in many cases, are, so much more! Vendors should not fight against Amazon and their marketing dollars. What they should focus on is providing outcomes and Conversational Experiences.
The Importance of Conversational Systems
The takeaway from this discussion is that Conversational Systems, in the Enterprise, need to support operational excellence and organizational resilience. Operational Excellence is a balance between exceptional experiences and operational costs. Adding AI into the discussion may help, but may also confuse (if not done properly). I would like to see Software Vendors focus on the outputs of the system and then highlight where AI is a better solution and provides a better experience for customers. This is topic is too important to become marginalized by leading with AI, when a really well-designed system might suffice. In a really good article from the founder and CTO at Pindrop, Martin Reddy highlights the importance of the human element. This nails it:
“Conversational experiences today can be quite simple and constrained. In order to move beyond these limitations, we will need to support higher fidelity conversations. ”
But the article also highlights that the discussion is much bigger than just voice and much bigger than just AI elements within conversations. At the end of the article, the importance of people is highlighted, which feels like a contradiction – Pindrop focuses exclusively on Voice Conversations. I need to make it abundantly clear that Artificial Intelligence has a place within the Conversations, CRM, and now Conversational CRM. Actually, there are many places where AI is going to be increasingly important – let’s just be cautious as to where AI fits and what role it is meant to play. There are some great companies, solving some really tough challenges (Healthcare, Contact center volumes, and more…). My purpose here is to help these Vendors to take a pragmatic approach to terms and their marketing efforts.
good to read from you again, Mitch!
A very interesting read Mitch. The key point is that consumers want to have meaningful conversations, on their channel(s) of choice. AI is just part of achieving this – it is never going to be able to handle the entire conversation. Gartner have also talked about this trend – we’ve looked at their definition in this blog post https://www.eptica.com/blog/why-brands-need-focus-conversational-cx-platforms
Apologies for the delayed responses – US Holiday travel and such…
Hi Thomas – and thanks!
Pauline, Hi and thank you. I am a fan of conversational experiences and conversational CX seems to fit. My approach is a bit more generic, as I view conversation experiences to take place within organizations as well, hence not all are customer (employee, community or member) as well.
The #AI bit is a discussion as well – my key point is that intelligence is more important to the actions and capability, the Artificial part will become more important, over time.