An End-User’s Story: Looking at a Real-Life Customer Experience Failure

We recently received a story. It was from a Cigna customer reaching out to their helpline and getting connected with a robot—we call these auto attendants—instead of a live person. We’ll save you the suspense… the interaction didn’t go well. Chances are, you’ve experienced something similar yourself. Often, these attendants don’t understand what you’re saying. They give you the wrong information. They’re hard for you to follow or understand. And if you’re a company that values your Customer Experience (CX), these solutions just aren’t cutting it.

At the same time, the technology exists for AI and automation tools to provide real value. But you have to use it right. We’ve spent a lot of time talking about what this looks like—here are some key points:

  • First, customers want to be able to find answers by themselves. Use AI to offer self-service options.
  • Second, you should be using AI to make important information available at your agents’ fingertips.
  • Third, be aware that there are things that AI typically doesn’t do well—like handling complicated requests or sounding natural. Don’t overly rely on it and let it derail your customer interactions.

That third point is critical. Even if your auto attendant gets it right most of the time, are you really willing to let even a portion of your customers have a bad experience? And are you getting the right insights and feedback to even know how often these conversations are frustrating your customers? These negative experiences are occurring every day and driving customers mad. Instead of just talking about the problem in hypotheticals, let’s look at a real, actual interaction of a typical end-user experience—and how you can make sure it doesn’t become your customer’s story.

When A Company’s Helpline Is Anything But Helpful

I recently had the opportunity to try out the helpline for Cigna’s home delivery pharmacy, and wow, was it a frustrating experience! 

I am the type of person who always tries to solve problems by myself before making a phone call for help. I hunted all through Cigna’s website for how to do what I wanted to do—piecemeal, as they have no “search” function— and couldn’t find anything to answer my question. They also had no chat function with an AI or live human to help offer clues. I was stuck with calling.

According to Microsoft, 90% of customers expect to be able to use self-service tools to address their questions. Whether it’s opening up your knowledge base to generative AI, implementing a chatbot on your website, or giving your agents omnichannel tools to respond to texts, chats, and inquiries through social media, your customers want to be able to reach you on their preferred channel. Be there for them.

As I expect these days, the call was answered promptly by an automated system. Again, as expected, it verified my phone number and identity to pull up my record. So far, so good.

This is actually a nice start. Integrations are available so that all of the information you have on your customers should be at your agents’ fingertips. When a customer calls in, you already have their account name, account number, purchase history, and other information. There’s no reason to ask for what you already know! Smart database integrations will lead to quicker and more pleasant interactions.

The problems started when the attendant asked what I wanted. I would guess that the system is programmed only for “ordinary” questions— but I could have found the answer to any ordinary question by using their website. Consequently, it did not understand my verbal reply, as it was outside of the expected list of queries.

Don’t miss this key point. Auto attendants are programmed to handle the common, simple questions. This isn’t a problem—those likely are the majority of your calls. But the difference between great customer service and a poor outcome with your customers is how you handle the rest of your inquiries. The easy answer? Provide a quick, clear way for your human agents to jump on these calls!

It then decided, completely on its own, to start reciting my most recent orders and telling me whether they were delivered, followed with, “Is this what you want?” No, you stupid AI, you’ve gone off on your own wild goose chase! I finally asked, when it eventually offered me a voice prompt again, to “speak with a human being.” It replied, “I hear your call for help,” which was a weird response. It then re-verified my information (why?) before connecting me to a person. 

All of the above was made worse by having poor voice generation, with odd word spacing, inflections, and pronunciations. It sounded like early voice generation from the 1980s. It was really obvious that it was a computer-generated voice, and it made it harder for me as the listener to parse and understand. There were also many long pauses, which were long enough that I wondered if I had been disconnected at times. It also didn’t tell me what it was doing (not even a, “Please wait while I complete your request”) so I was left wondering if it had heard me at all, let alone understood what I asked for.

The technology is there for conversational AI to speak naturally and understand fluently. So why isn’t this your experience when you call in? They’re using outdated technology! You can’t implement these types of AI-driven or automated solutions once and then leave them alone. These technologies need to be frequently re-evaluated and assessed. Using technology to help you create great customer experiences is a continuous process, which is why you need to have an on-going strategic partner that can help you meet your customers’ evolving expectations.

After speaking with the live human being (and finding out that there actually was no way to achieve what I wanted to do), I offered her feedback on making my experience better. She asked me to stay on the line and complete a survey, which, she said, would give me the opportunity to officially put into the system my recommendation. So, I waited.

The survey (with a different auto attendant, and better voice generation) asked me four questions about the call (without the nuance of distinguishing between their AI and their human). There was no opportunity to offer feedback in my own words about either the call experience or my original question. It was only, “rate this general aspect of the call on a scale of one to five”. The topics they asked about were worded in such a way as to push me toward higher numbers than they deserved, because of the tricky wording. (Note: the only thing I scored highly was the live agent, as she was perfectly fine, and she had nothing to do with choosing the AI or setting company policies.) From my previous experience designing surveys, I suspect that the four questions they chose were not even offering useful data for evaluating the calls; it felt much more like a tool to collect (inaccurate) marketing data.

If great customer service is important to you, you should be paying attention to customer interactions and trying to find the holes in your CX so you can fix them. This means getting helpful, constructive feedback. How can you make a customer’s frustrating experience right if you don’t even know if they’ve had a bad interaction? Well-constructed surveys are part of this. You can also implement AI technology, like 8×8’s Conversation IQ, to get real-time feedback on what’s happening on these calls, complete with sentiment analysis that makes sure any frustrating interaction is on your radar.

I think this experience has probably been repeated by many customers, as most people I know would only use the phone-in helpline as a last resort. And with no way to offer genuine feedback, Cigna is never going to improve its system.

Cigna, if you’re reading this, we can help you out.

Stop looking for quick ways to save money and start looking at these interactions from a customer’s point of view. Think about how you can give your customers the self-service options they’re looking for on the channels they prefer. Always have a human agent available when the question goes beyond what a robot is capable of handling. And know what your customers’ experiences are and where you’re missing the mark.

That’s where choosing a partner like Vertical can deliver an invaluable return. Our industry-leading lifetime Net Promoter Score (NPS) of 83 shows that we know what we’re talking about when it comes to customer service. But it goes beyond the numbers. It’s all about our customer-first approach, which we call the Vertical Difference. We are your strategic partner as you build out your communications. Our experts work with you to design and implement innovative solutions that will grow your bottom line. Our technicians and project managers are on-site for the white-glove installation and implementation. And, crucially, we stay with you for ongoing service and support. We’ll make sure your system is always up-to-date, meeting your needs, and future-forward. When you trust Vertical, you are ensuring a story like this won’t be coming from a customer of yours.