Businesses are moving their “bot strategy” from vision to reality at unprecedented speeds. Now they have help from solution providers with development platforms that fulfill on the promise of applying Machine Learning and Deep Neural Networking to the hard tasks involved, first, in identifying high-impact use cases for bots and, second, to build conversational models that create the shortest path from intent recognition to resolution. Successful deployment of these solutions can save large companies millions of dollars in staff time (both internal and 3rd Party professional services) and tens of millions in labor costs as chatbots or virtual agents gain acceptance from customers and employees alike.

The need for speed in bot deployment was initially satisfied by solution providers with libraries of pre-configured conversation models, handling the categories of calls and intents for specific vertical industries. When choosing a solution provider in the world of “Answerbots”, decision makers simply asked: What works out of the box?” and “How quickly can you install a bot to answer my customers most frequently asked questions?” In many cases, this remains a viable strategy for getting started. A generation of customers are delighted to find a resource that enables them to use their own words, or press pre-defined action buttons, to get their questions answered without bothering other humans.

But experienced businesses want more as they look for tools that help them cull through voluminous amounts of conversational input to discover the categories and intents that define the next big use case and then put tools into the hands of developers to craft conversational models make engagement with customers more pleasant and expeditious. There is a longer-term requirement of applying elements of AI, specifically Natural Language Processing, Machine Learning, Predictive Analytics and “Cognition”, to support a sustained, conversational engagement 24/7/365, spanning a wide variety of topics or objectives and employing each customer’s device of choice.

Nuance Project Pathfinder: Speeding Conversation Development

On February 5, Nuance Communications introduced Project Pathfinder at its Customer Experience Summit. In a subsequent briefing with Opus Research, Nuance’s Paul Tepper walked us through a demonstration of the product’s capabilities, noting that conversational dialog designers at a select group of strategic clients were already putting Pathfinder through its paces. Their mission is to use the platform to break through today’s status-quo for bot design, which is labor-intensive, time-consuming and error-prone. As a “data-driven design tool” Pathfinder ingests a voluminous number of chat transcripts, “tags” them based on the intent(s) of the chat, groups them by topic or category and lays the foundation for conversation developers to use a “Visio-like tool” to design conversations.

The name “Pathfinder” captures the overall intent of the platform which is to find the straight path, amid multiple turns and topic changes, from a recognized intent to ultimate task completion. It will be made generally available as a product during the summer of 2019.

Discourse.ai’s Cognition Technology

You’ll see a new category of solution evolving here as Nuance is preceded in the marketplace by the product offering of Discourse.ai, a start-up that is gaining traction among strategic enterprises in Financial Services, Telecommunications and E-commerce as it defines solutions for “conversational automation.” Cognition’s Machine Learning resources perform the following functions:

Cognition is already generally available in the market. Based on its experience with strategic clients, Discourse.ai sees Cognition speeding the “time to value” dramatically, with the potential to reduce development costs by as much as 90%.

From Google’s Area 120 Comes Chatbase

To further legitimize this category of product, Google has launched Chatbase, a cloud-based service “that replaces the risky status quo approach with a data-driven one based on Google’s world-class machine learning and search capabilities.” Chatbase was launched out of an internal incubator called Area 120 where it uncovered significant organic demand among bot developers and Google was able to refine the product and services based on “analyzing hundreds of thousands of bots and billions of messages in our first 18 months of existence.”

While the nomenclature is slightly different, Chatbase uses Machine Learning to analyze conversations with millions of bots, identifies the underlying customer issues which it calls “drivers” and then culling out multiple intents within those issues. Once again the time savings, and therefore cost savings, are impressive. Google puts it, in some cases it reduced development times by a factor of 10x, while fitting 99% of conversations into recognized intents.

Conversation Accelerators: A New Category of Platforms

Conversation Acceleration Platforms (CAPs) comprise only part of the full stack of Conversational Commerce solutions. Yet they are in a part of the value chain where they have proven economic value which can be the basis of calculating return on investment (ROI). Hiring Conversational Dialog Designers, internally or through 3rd Party integrators, is expensive. Reducing the amount of time it takes for them to do their jobs provides immediate savings. Plus there is the added value of bringing new bots or intelligent assistants into the conversational path with customers more quickly and effectively.

Categories: Intelligent Assistants