Have AI chatbots come of age?

Conversations with chatbots are getting a lot more interesting. Fueled by artificial intelligence and natural language processing, chatbot technology has taken leaps forward over the past few years.

“Chatbots are now a critical component in the customer omnichannel experience,” said Rob Dunlap, Associate Partner – Cognitive & Analytics Practice Leader with IBM at a recent CanadianCIO Virtual Roundtable.

In the past, the introduction of chatbots hasn’t always delivered better customer service.  However, the technology is changing rapidly to meet consumer needs, said Dunlap.  “Personalization of the experience where the agent has knowledge about you is an important part of the way the technology is evolving,” he said.

The adoption of virtual assistant technology has been steadily increasing over the last year. “If we want to be competitive in the omnichannel, then we will need these tools,” said one of the roundtable participants.

Improving customer service at scale

Most of the roundtable participants said their organizations are just starting to look at chatbot technology. One said that his management is hesitant to move forward because of a desire to have a real person answer every call. Another concern was that it might take too long to train the machine learning in the system.

User experience is critical, said Dunlap. “Research shows that, when it’s done well, leading firms in the virtual agent space can increase customer satisfaction,” he said.  “They’ve done so by automating an end-to-end delightful experience in the bot.”  For example, he noted that Purolator uses IBM’s Watson Assistant to provide the status on package delivery for customers. This has improved the company’s ability to deal with a big increase in customer service inquiries over the past 15 months, he said.

Chatbots do not replace all human interactions. The process can start with the bot, but if it can’t answer the question, the call can go to a human agent through the same interface, explained Dunlap.

The latest AI chatbot packages are intended to be more accessible and easier to implement, Dunlap said. “It’s already set up for natural language,” he said. “Our clients only need to provide 15 to 25 use cases to set up the system, so you don’t get stuck at the beginning.”  The system also collects better data on customer questions, so that it can adapt and learn to deal with additional business cases.

Tips for chatbot success

The best way to get value out of a chatbot is to choose the right use cases, said Dunlap.  “It’s really important to be fact-based in determining the use cases that can be automated from end-to-end,” he said. “That’s where the real value is.” Defining goals and metrics up front is another critical step, so that organizations can measure progress and continuously improve.

To ensure that the chatbot is well constructed, the team should engage the people that understand the process, such as call centre agents. “This is an important differentiator that makes the chatbot more successful,” said Dunlap. When organizations collaborate with the subject matter experts, they become champions of the technology and can help drive adoption in the enterprise.

“We believe it’s a partnership between human authors and AI,” said Dunlap.  “This is how you build a phenomenal automated experience.”