Navigating the Conversational AI landscape

The conversational AI space remains challenging to explore with over 350 players. As I prepare the next versions of my SalesTech and CX landscapes, I wanted to share my latest taxonomy to navigate this crowded space.

The dichotomy between unsupervised and supervised learning continues to serve as a useful framework for customer service and support solutions.

I have termed unsupervised solutions that leverage technologies like semantic search or question-answering as "answer bots" - borrowing the term from Zendesk after its 2017 product launch. These transform the search experience by only surfacing the most relevant results. So far, brands have largely purchased them for customer support and website help centers with loose integration to contact centers. With generative AI, these solutions are gaining a significant upgrade in capabilities. I'm renaming this category "Generative Answers & AI Search" to highlight the new possibilities enabled by generative models.

Supervised solutions bifurcate into applications and platforms. Prebuilt applications target specific use cases (known as "intents") - categorized as "Conversational Commerce" for sales or "Intelligent Virtual Agents (IVAs)" for customer service. Mastering a seamless voice experience remains challenging, prompting me to spotlight “Conversational IVR” as a distinct category.

Platforms manifest in different forms. I term "toolkits" the components targeting developers - offered by cloud hyperscalers and select others like AWS, Google, IBM, and Microsoft. Toolkits embed their proprietary AI technologies, facilitating the creation of "composable" applications. "Toolsets" come from specialized providers like Cognigy, Kore.AI, and OneReach - featuring low-code application development, intent creation/management, and instrumentation/testing tools. They increasingly enable customers to choose their preferred AI technologies, like NLP engines.

Conversational AI is expanding beyond purely inbound use cases into more outbound applications. Interactive notifications now often incorporate conversational capabilities - positioning these for a pivotal proactive service role (“Conversational Engagement”). In sales, Conversational AI unites with Generative AI to create digital workers who can work as sales assistants. They can engage in a conversation to qualify incoming leads before passing them to a sales rep or engage new prospects based on target Ideal Customer Profiles (ICPs) and personas - applications I dub "Autonomous SDRs" and "Autonomous BDRs".

I must acknowledge this taxonomy falls short in addressing important questions from industry practitioners, including:

1. What skills and effort are required to drive results out of the solution?

2. How can brands safeguard quality experiences?

3. How much effort is required to make changes and expand the range of supported questions and intents?

Such topics are best explored through experience sharing and peeling back the curtain on underlying technologies.