Feb 12, 2019 When Humans Fail, AI Can Rescue Customer Experience
I have talked to many friends and colleagues who complain about people who don’t return calls, emails, and texts—either at all, or when they say they will.
These are typically business-to-business or business-to-consumer interactions (vs. personal calls), where the supplier tells a customer: “Let me check on that, and I’ll (call/text/email) you back (in 10 minutes, an hour, by the end of the day).”
As the clock ticks away, there is no response. In fact, the next day, still no response. And the customer now has to reach out again (and maybe again and again) to get what he or she needs. It’s an incredibly frustrating experience, particularly when you’re the customer and you’re trying to buy something or want service on something you’ve already bought!
In researching Digital Customer Experience for several years, I have yet to see some innovation in this very common and basic failure of humans—to follow through when and how they say they will.
The Value of AI
Here is where artificial intelligence (AI) and its derivatives can help resolve the problem. By using Natural Language Processing in written communication, and speech recognition in verbal communication, AI engines can either prompt someone to return a call, text, or email on time. Or, they can deliver a response by default if the human fails—keeping the customer informed, updated, and not feeling left behind.
Take an example of a regional manager who fosters numerous customer relationships, resulting in significant multi-tasking. He drives a lot to customer sites, so phone is the common method of communication. One of his customers escalates an urgent problem to him: An order that was supposed to be delivered the previous day never shipped and he needs it for a high-value installation already scheduled.
The manager says: “I’ll get this resolved and call you back within 30 minutes to provide an update—promise.”
Within those 30 minutes, another call comes in and distracts him from resolving the original problem. The day ends, and he goes home—never to call the customer back or resolve the problem. So much for promises. It’s not too hard to imagine the customer’s frustration.
If AI were part of this manager’s technology toolkit, the outcome could be much different. Suppose his phone rings in 30 minutes, and a bot says: “It’s time to call Customer X back to update him on his shipping problem. Say ‘call’ to ring his phone.” At that point, any good customer rep will call the customer and let him know he’s still working on the problem and needs another 30 minutes. At that point, the AI engine understands the next step and reminds the regional manager again in 30 minutes if he doesn’t initiate the call himself.
AI Becomes a True Assistant
Other options might include default settings. Suppose the manager is on the phone with someone else at the time he is supposed to call his customer back. The bot can either text, email, or call the customer with a notification that the regional manager is still working on the issue.
Similarly, if someone texts, webchats, or emails the manager about the problem, a NLP chatbot can make sure the interaction stays on track. And in some cases, it can check on the order, update the customer, and resolve the issue entirely for the busy, multitasking human manager. For example, maybe weather issues caused the order delay, but no one updated the customer. Or maybe the credit card number was wrong, which delayed shipment because no one corrected it with the customer.
The key point is that people want to manage to expectations. They want updates on schedule. If there are problems with deadlines, they want to know proactively so they can regroup—and not have to chase someone down to get answers. Often, people avoid those difficult calls because they a.) don’t want to deal with the conflict or an upset customer at that moment; b.) they simply forget (we’re all human).
Muting this common frustration through the use of AI would go a long way in improving the overall customer experience.