As discussed in the first blog of this series, conversational AI is a breakthrough in human-computer interaction, as it enables a natural language exchange. But for first-timers, there’s a steep learning curve and a lot of trial and error to implementing conversational solutions that ultimately exceed expectations.
After more than 1,000 client discussions, hundreds of opportunity assessments, dozens of vendor evaluations and briefings, and an expansive portfolio of proofs of concept, pilots and deployments, we’ve learned a lot about what distinguishes successful conversational AI solutions from those soon abandoned. Here are the most important lessons learned.
- Align business and IT. We’ve often seen major issues develop with projects that start from an isolated unit with no alignment between business and IT. Conversational AI projects require full support from internal IT departments, as integration between the cognitive agent platform and front- and back-end systems needs cooperation and approval from IT product owners. Additionally, projects initiated solely by IT departments risk missing business strategy alignment and the right set of objectives to deploy an end-to-end solution with clear benefits.
Start with customer journey mapping. Previous projects have shown that businesses often lack a clear picture of what the customer wants when it comes to cognitive agents, and are more focused on day-to-day operations or anecdotal insight. Businesses that develop a cognitive agent without completing a proper customer journey map stand a high chance of designing the wrong conversations that fail on either customer experience expectations or don’t address the core issues of customer needs.
- Favor choice over one-size-fits-all. Early on, we learned that businesses need choices to be able to accommodate the many products and platforms customers wish to use. Our practice partners with the leading cloud AI framework providers, including Amazon (Alexa Advisory Council and Alexa digital agency partner), Google (Google Cloud Premier Partner), Microsoft (MPG Gold Partner) and IBM (Platinum Partner and Advisory Board Member for Cloud, Cognitive and Watson). We also partner with specialized point solutions companies, including Artificial Solutions (Teneo Partnership Program) and Conversable. With access to such a broad array of best-in-class technology, we can deliver across all the channels, languages and platforms through which our clients and their customers are engaging. For our solution architects and developers, these relationships give us access to partner labs, advanced insight into roadmaps, early access to product betas and influence on future enhancements and features.
- Incorporate the conversational design basics. From past engagements, we’ve developed a template for conversational design that we continue to refine. Some of the conversation basics we incorporate include three-second wait for reply, short sentences, inclusion of videos/images, drop-down choices, suggestions, small talk, a balance between politeness and a functional focus, an escape option when the wrong selection is chosen, optimized screen size, clarification of the scope of the cognitive agent at the beginning, and an explanation provided when the cognitive agent doesn’t understand a question.
- Understand the complexity of selecting a natural language solution. It’s misleading to base the choice of natural language understanding for a conversational AI solution solely on its upfront costs. Among other customer-specific requirements, a typical enterprise solution will need to consider channels, security, hosting, languages, conversational analytics and customizations to determine the best fit. To make the most informed decision, businesses need to consider all the implications of the various factors that directly or indirectly impact the cost of the total solution.
- Use Agile, not waterfall, delivery. As conversational AI is a new topic to most businesses, they often don’t know what’s expected from them in designing a cognitive agent solution. For example, if key information about products or services is missed in the requirements analysis, it means that conversations need to be redesigned. These iterations can only be supported with an Agile delivery method. Customer validation during each Agile sprint becomes critical as well, along with frequently requesting and incorporating feedback into subsequent design phases.
- Plan for regression. Because of the risk of regression impact on existing intents, it can be costly to accommodate even seemingly minor changes to the conversations in the later stages of the lifecycle. This is because even small changes in dialogs will require retesting of the full conversations for intent conflicts. To address this challenge, our conversational AI lab and R&D team has developed automated testing tools specifically for regression testing of dialogs across all the major vendor frameworks and products.
- Be prepared for continuous benchmarking of conversational AI technologies. Conversational AI services are still at a nascent stage of evolution. While multiple natural language understanding services exist, it’s critical to appreciate the strengths and considerations of each option as it evolves. Benchmarking natural language understanding services, therefore, is a critical activity for success in this space. Our practice has a dedicated lab for benchmarking market products on an ongoing basis to ensure we’re providing the most current recommendations to our clients.
Think big, start small. Chose use cases and projects carefully. Because each business has its own dynamics, conversational AI solutions are not a “copy/paste” from one client to another. A specific learning curve is required to be effective and efficient in the implementation of a cognitive agent.
Preparing Now for the Future
As I’ll discuss in the next blog post, conversational AI is a fast-growing specialization that is only now getting started, and businesses need to develop a flexible strategy that takes into account the expected developments in this area over the next two years. At the same time, now is the time for digital change leaders to help their companies bridge to the conversational AI future, as doing nothing is less of an option with every passing day.