3 Questions to Ask Your Conversational AI Vendor

For customer service, conversational AI bots are a necessary tool to handle an influx of customer queries and to deflect easy questions that humans shouldn’t have to answer. And instances of conversational AI are only increasing, considering that 74 percent of organizations say conversational assistants are a key enabler of the company’s business and customer engagement strategy.

If you’re looking to implement a chatbot for your own organization, you’ll obviously want to carefully consider your options. Here are three questions to ask when you’re making your evaluations.

Question 1: What are the bot’s use cases?

You need to meet customers where they are, and in today’s world, your customers are online. Even in 2015, Forrester found the use of help or FAQs on a company’s website increased from 67 percent in 2012 to 81 percent among U.S. online adults. Customers would much rather look for an answer online than escalate their request to a call center. Enabling customers with a conversational AI bot—whose natural language processing allows them to ask their questions in everyday syntax—gives them the information they’re looking for with a minimal amount of friction, leading to a good customer experience and increased satisfaction.

It’s also important to consider whether the bot can be used across the organization both internally to increase employee productivity and externally to increase customer satisfaction. A bot can help customer service teams save time by deflecting tickets, therefore increasing the volume of support requests they can handle. It can also save employees time by delivering the answers they need without having to perform an extensive search—for example, locating benefits information, retrieving a specific document that is “lost” on the shared drive, or discovering how much PTO time they have. A conversational AI bot with an internal and an external use case will deliver much more value over time than one that can solely be used with customers.

Question 2: How much time can it save?

A big value proposition of conversational AI bots is their time-saving abilities. As noted, a bot can help customer service teams save time by deflecting tickets, hence increasing the volume of support requests they can handle. Think about it: If a bot could deflect 75 percent of all helpdesk tickets, customer service reps internally or externally would have a huge chunk of time to dedicate to other projects.

In an internal use case, the power of a bot truly shines. Consider how often people are interrupted by colleagues with questions about where to find a certain document or a certain process. Researchers at the University of California, Irvine, found the typical office worker is interrupted every three minutes and five seconds. And with each interruption, it can take 23 minutes and 15 seconds to get back to where the worker left off after interruption. Those interruptions eat up 28 billion wasted hours a year, at a loss of almost $1 trillion to the U.S. economy. Let’s go back to our previous example, with 75 percent of all questions being answered by a bot. If a conversational AI bot is implemented, the internal team won’t be interrupted as frequently and can focus on cross-selling and other tasks that require higher-level thinking.

Question 3: Does it continuously learn from organizational knowledge?

A bot has to pull information from somewhere in order to provide the right results, so you’ll want to ask any vendor what kind of knowledge base the bot will pull from. This can be the most arduous part of an implementation, so weeding out vendors with a weak experience in this area is critical. Once the bot has centralized knowledge to pull from, it should have some way to evaluate whether the answers it is delivering are the right ones, so that it continues to get smarter. This can be as simple as a post-answer thumbs-up or thumbs-down, giving the technology the feedback it needs to improve.

Finally, at some point, the conversational AI bot will come across a question it cannot answer and a human will need to intervene. Pay attention to how the bot escalates questions, either by looping in a human via an automated workflow or incorporating a human attendant in live chat.

As conversational AI bots proliferate, it can be hard to know how to choose the right technology for your organization. Armed with the questions above, you’ll make a well-informed decision and reap the benefits of a great bot.

David Karandish is founder and CEO of Capacity, an enterprise artificial intelligence company building a secure, AI-native support automation platform to help teams do their best work—and save time and money. Before founding Capacity, Karandish was the CEO of Answers Corporation. He and his business partner started the parent company of Answers in 2006 and sold the company to a private equity firm in 2014 for north of $900 million.