×

blog

3 Lessons for Chatbot Initiatives

Chatbot Initiatives

Stay Connected

Blog updates via Email

The use of chatbots in the customer care space is a hot topic for contact center leaders. Chatbots are an essential AI program that contact centers are utilizing for simple customer tasks and knowledgebases. According to IBM, bots can respond to around 80 percent of routine interactions, allowing associates to tackle complex and emotive questions. But organizations need to understand the consequences of bot technology before implementing it head on.

A Customer Contact Week (CCW) report, “The State of Chatbots” by Brian Cantor, outlined the dangers of applying bots without consideration and the questions organizations need to ask before implementing chatbot programs. Here are three key lessons we took from this report.

Lesson #1: Don’t rush

As more organizations use AI technology, many leaders believe it’s progressive to have chatbots on deck. But rushing bot technology just because it’s the shiny object can be reckless and irresponsible.

Cantor explained that chatbots do not automatically ensure a functional and seamless customer service environment. A faulty chatbot can poorly communicate and provide irrelevant information. This can completely contradict the whole point of bot’s capabilities to provide automated, 24/7 self-service for simple tasks, instead becoming a frustrating obstacle for users. Consider the level of discomfort customers are going through when they are reaching customer service, if a bot operates poorly that easy request for store hours can become a frustrating experience.

The report urged contact centers to listen closely to skeptics of bot technology. Yes, it is difficult to go against what is popular in the tech space, but listening to those who fully understand bot technology and their applications may help prevent many of the common mistakes these tools face, such as customer dead ends, lack of knowledge, and misunderstanding of requests. It’s not most important to be the first contact center to use chatbots; it’s most important to be one of the first to use them correctly.

Lesson #2: Place bots in the right spot

Contact centers need to understand a bot’s purpose and the value it creates in everyday interactions. Today’s customers seek help on the channels they want, when they want, as evident with Messenger apps and social media chat. When tech-savvy people scroll for help, the bot needs to be placed in the most effective and useful channels.

But the report stated that bots are usually siloed in live chat or messaging without any relevance beyond self-service. This creates a disconnect for live associates and doesn’t ensure a smooth transition to them. Starting a queue with a bot then having to restate the problem over again to a live associate is frustrating. This can stem from companies who internally make the decision of where to place the bots instead of listening to the voice of the customer.

Cantor suggested that contact centers run a “customer centricity test” on their bots to analyze if AI on this channel makes sense. That means reconsidering if the bot can handle the complexity of the questions asked, if the conversation is natural, and if it provides the right kind of personalization for the task. Contact centers do not need to throw a bot at every issue. They should consider the right bot for the right job. Chatbots cannot be a vanity project, it needs to have substance, Cantor said.

Lesson #3: Think of the people

When implementing chatbots, leaders need to consider not only how they will affect the customer experience, but also the associate experience. Poorly designed chatbots create bad experiences that end up on a human associate’s front door. This can place them in an unfair position where the customer may already be annoyed from his or her failed bot interaction.

Remember, bots can never replicate the empathy and care needed for complex and often stressful customer issues. Instead, place a bot to solve simple tasks, and use people take care of the more valuable interactions. Leaving too many issues on a bot’s plate can over-complicate things and make its implication into certain areas useless.

The report urged readers to reconsider the bot’s relationship with humans. Bots can take a step back and become internal workers within a contact center by using AI to assist humans with internal processes and updating data and knowledgebases. It doesn’t have always be customer facing.

In addition, consider how introducing bots will affect the workflow your associates already handle, establish a secure escalation between bots and humans, and train your workforce on how to handle poor bot experiences if the need arises.

Cantor’s report is not meant to discourage the use of chatbots. Instead it’s a call to action for contact centers to reconsider the repercussions of poor chatbot implications. Chatbot technology is far from perfect, but if used correctly it can be an invaluable customer experience tool.