Customer Service in an Age of AI

Customer service is in decline. In my opinion, the problem is worse with large organizations.

I skimmed the surface in my last post “The Robots are Already in Control (Part One)“.

Companies that are primarily internet-based are the worst. One of my major pet peeves is a corporate website that has no phone numbers. For many, at best the customer service phone line is almost unfindable – buried four clicks into the site. I’ve even gone so far as to phone the “company” side of the business (the side that deals with shareholders), because there was no number available for customers to use!

21st Century Customer Service: Sapping your Time, Saving them Money

We know why this is. Organizations think they are saving money by directing traffic to FAQs, crowd-sourced Q&A or “contact us” forms. Why pay for service when you can make the customers do it themselves? I cringe thinking about the senior citizens I see trying to bag their own groceries.

I remember the days when we knew our branch manager at the local bank. It was great to have someone who knew you, who knew their job and who was willing to go the extra mile to get things done for you. That is the beauty of a well-networked human. If they don’t know the answer, they know someone who does, and will pick up the phone to ask them for help.

With the current customer service models that they’ve grown up with, I fear that the next generations of front-line service staff have never experienced this. Maybe they’ll never know what great customer service looks like. Put your terrific customer service experiences in the Comments!

I raise this because, over lunch recently, a friend told me of her experiences going to a local bank in the town she had just moved to. Exasperated with the bank website, she visited a branch to see if they could connect her to the Toronto office. After a lame attempt at looking through their local information sources, the young staff member did a google search for the branch info! My friend was shocked because OF COURSE she had already tried that herself!

These days, we likely only pick up the phone to call a service line if our problem is complicated. If we can’t solve it with the online tools, we need to speak to a human! We all expect to wait long periods for any service agent to pick up the phone.

AI customer service frustration

Additionally, we know that in lots of companies, the front line staff are under-resourced.

They likely haven’t been given permissions in the system to deal with our most serious complaints.

Escalating to a manager may never occur, and if it does, may take days for a response.

Enter AI

Now layer AI into this scenario.

AI may answer the line faster. It will be able to answer simple questions. Presumably, AI will be using the same datasets and information sources as we were given when fumbling around on their website. But will it be able to do much more than that? How is that useful to most customers who phone the line or get on a chat with them?

Will AI be able to manage the layers of complexity that it will be presented with? Humans have complicated lives. Serious responsibilities. Relationship, financial and time constraints. Emotions and varying degrees of patience. Language barriers… the list goes on and on.

I think that robots will never be able to finesse this complexity; at least not in my lifetime.

Undoubtedly, an AI chat bot for an airline will not know how to compassionately listen and respond to an hysterical customer. It is unlikely to make provision for a last-minute seat request for a person’s whose mother passed away abroad. (Interestingly, fours days after writing this post, the following news article flew by my inbox:

Nick Roberts, “Air Canada must pay refund promised by AI chatbot, tribunal rules” (Yahoo News, Feb 18, 2024), online:
https://ca.news.yahoo.com/air-canada-must-pay-refund-040527669.html )

So I stand corrected, and now wonder whether this decision may actually force companies to put guardrails around their chatbots so that they do not make promises the company doesn’t want to keep! (My colleague Pascale Chapdelaine recently argued over dinner that she thinks there are plenty of existing tools in the judge’s toolbox to deal with the challenges that lie ahead. So perhaps the Air Canada ruling is a taste of things to come!)

I still think that a bot will not have human relationship advantages, such as having a friend in the business elsewhere from whom they can call in favours.

AI Can’t Help if the Data’s a Mess

Another example that plagued me for the past couple of years was the complete obscuring of contact information for the Ontario Courts. The old website that still exists and is dominant when you google for them, for awhile only contained addresses. It did not provide telephone numbers. Instead, a person somehow had to know to go to the (not cross-linked) Ministry website for phone numbers. Thankfully, they were conducting a user survey last year, enough of us gave them a scolding, and the problem appears to have been fixed up now.

Ultimately, how is AI going to help if the basic data sources are a mess?

AI is being built on top of the customer service mess that companies are already serving up. I question whether a bot will be able to smooth over this mess for the company the way a human could.

As humans we’ve developed a variety of methods of getting sympathy from others. How many service agents have you spoken to that readily admit the company has challenges. They do this so that we’ll recognize that the issue is not their doing, and give them a pass. I don’t think I’d be inclined to so readily give a chatbot a pass.

We all know a human customer service agent who really went the extra mile for us. Someone who was compassionate and a good listener. Sometimes, that makes a huge difference.

Will AI Compound the Problem?

Surely the problems we are seeing with austerity – driven customer service decline will only get worse.

Companies will lay off human service agents, thinking that the bots have got things covered.

Will the bots simply talk right over us, as I described in my last blog post “The Robots are Already in Control (Part Two)“?

Significantly, will the bots have authority in the system to fix an invoice, issue a credit, re-issue a new ticket, etc? From my experience, many front-line human customer service agents haven’t been given this authority, so it’s doubtful that an AI chatbot will.

The upshot of frontline staff (human or not) without sufficient authority in the system, is the same old cycle: customer must request (multiple times) and / or wait on hold again to have our question escalated to a supervisor. Hours are spent finally getting connected to a person with sufficient authority to do something useful for us.

Hopefully these questions will be food for thought! At least they’ll give us a place where we can vent with other humans!

Post Script
-After releasing this post, a reader contacted me. In the email, I was reminded that this entire situation laid out above, is compounded for people who have not adopted technology (for whatever reason).

Imagine trying to get customer service these days with only a land line (no cell phone and limited facility with technology)!

I was grateful for the wake-up call and encourage public libraries to continue offering technology skills training for the public. Those skills are important now more than ever before…

Comments are closed.