I noticed an article on MacRumors that Apple is getting ready to roll out it’s “AI-powered Support Assistant.” Apparently Apple’s only willing to take the risk with tech support and not customer support. That’s where money might exchange bank accounts.
This is the way the world is spinning at the moment. My prediction is that companies will ratchet back these cost saving moves but not until the realize what it costs in customer loyalty. We have to go through the suck, before we get to the less suck.
I base that prediciton on my experiences spending the better part of this Summer and past Spring helping clients and a few relatives, some elderly, work to cut back on expenses and take care of some warranty repairs for appliances. (Hint: Ditiching warranty services is one way to cut montly costs.)
Timing is everything. My efforts coincided with several of these companies caught in the process of switching to AI solutions for customer and tech support. To be direct, it was a mess. For me, the customer, and also for the employees left holding the bag on the other end of these crazy corporate strategies.
No one would argue that most customer and technical support systems were in good shape previous to these kind of moves. They were indeed ripe for re-inventing and they have always been easy marks for the bean counters to cut corners.
Chatbots had already begun to proliferate, descended from automated phone trees like some form of inverted evolution that only a sadist could love. No one would ever conclude that they had been intelligently designed. My previous experiences were always hit and miss, but at least I understood that if I followed the steps I could eventually reach someone on a phone or live chat. During these transitons that became nigh on near impossible.
Prior to one of the companies I worked making the switch, it had become obvious that they had abandoned call centers and let their agents work from home. I kid you not, I spent one phone call with a barking dog in the background of whatever small quarters this rep was in, and another with two small children fighting with each other in the background of another call.
Let me give you one example from my recent experience with one company undergoing a transiton to AI tech and customer support.
Sears Home Services
Sears Home Services has been a popular home warranty service for many throughout the years. It’s been apparent for a while that services like these have lost their luster and some have devolved into scams. But for folks of certain generations they were always in the monthly budget. In my experience with one of my clients, prior to their switchover, it worked about as well as it was advertised to work. You made a call, talked with an agent, set up an appointment, a technician arrived to check things out, ordered parts if necessary, rescheduled the appointment, and then came back to effect the repair or replacement.
The first repair I assisted with happened just like I described above. The entire process from first call to final signoff on the repair took 10 days.
The second call not so much.
I guessed things were in trouble when the phone number prominently displayed on the webpage would not yield a method of speaking with an agent, but kept pushing me to their chatbot. The chatbot had limited options that you could select. It did not have a way to enter any request beyond those options. My guess is the company didn’t want customers speaking to an agent.
Undaunted, I did the usual online searching for phone numbers and finally stumbled on to a Reddit thread where users experiencing the same problems were reporting phone numbers that worked. Until they didn’t. It was a cat and mouse game of dialing a phone number before it was changed or taken out of service. That lasted about a week.
Finally I succumbed and scheduled an appointment via the chatbot. That chatbot sure was happy. I was told the support technician would contact me the night before to schedule a more specific appointment window. Things were looking up. That call never came. The morning of the appointment I got a text telling me the appointment was scheduled between 8 and 5pm. In my previous experiences these appointments were scheduled in four hour windows.
On the afternoon of the repair, at 5:30pm I got a phone call from a support person saying he would be their in a half-hour. He showed up 45 minutes later. He diagnosed the issue, said he would order parts and they would be shipped directly to me. This was a change I’ll describe later. He then said he needed to go sit in his truck and work with the system to get the parts ordered and that would take a half-hour or so becuase the system was constantly kicking him out mid-order and he would have to start over.
When he came back to finalize everything for that appointment he reiterated that this new system was screwed up. I told him I wasn’t a fan either. We had a good conversation.
Previously, according to him, each night technicians like him would get a list of their appointments for the next day. They would go to a drop ship pickup location for any parts for that day’s appointments, head out and do the work. Now the system sent parts directly to the address of the repair and required the customer to reschedule the next appointment once the parts arrived. Because customers were not receiving notifications of these part shipments, the parts would just show up, often get returned, delaying the inevitable service.
Most striking in his telling was the fact that he no longer received the next day’s list of appointment the night before. Only the first one for the day. Once that was completed he would receive his next appointment. Again, according to him, the previous system in place for dispatching personnel was pretty good at scheduling the next day’s appointments with minimal travel time between appointments. Now with the new system, he would find himself traveling between appointments, often located far apart, more than he would working the repairs.
To begin to bring this ancedote about this repair to a close, suffice it to say that the entire episode from first call to eventual repair has not yet been completed. It begin on May 20. Once it was determined that the appliance needed to be replaced, not repaired (that took two appointments,) a new one was ordered. Then canceled by Sears Home Repair’s system. Then re-ordered.
On the night before the scheduled delivery I got a call that it would be delivered the next day. That morning I got a call saying the delivery would take place that morning. Ten minutes later I got a call saying that the appliance had not been delivered to the delivery driver. (It was too large for a home delivery.) This necessitated canceling the order, re-ordering and a repeat of the process.
Somewhere in the middle of the process phone calls with agents became possible again. Progress? Not quite.
After several discussions with several agents I discovered that they were as challenged as their customers were by the new system. I’m sure you’ve played the escalation game to get to a supervisor and I played that game here as well. At last count I’m waiting on approximately 7 phone calls from supervisors that never came in.
Eventually the appliance was delivered to the delivery person. But there was no record in the system of an appointment to install the new one and remove the old one as per the warranty.
It took a month and two more canceled appointments for installation before Sears finally gave up and said a check would be forthcoming to cover the installation costs. That check has yet to arrive. (That’s why I consider the transaction not completed.) We did get the appliance installed on our own.
But of course, we did receive the usual survey request asking us to rate how well Sears Home Repair Services performed. Once the check arrives and clears the bank, my client will be cancelling that warranty. That will be our response.
Xfinity
One bill I was working on reducing for another client was Xfinity. I probably don’t need to say more if you’ve ever dealt with that company. But I will. I simply wanted to call and discuss plan options for reducing the bill.
No where was I able to get through to a live person. I was sent links to chatbots that would promise a live person, if I followed along, but those calls never connected.
I eventually took a trip to an Xfinity store and got the info and made adjustments to my client’s bill in person. The staff member at that store told me that they were seeing more foot traffic because the online and phone systems were such a mess.
Back to Apple
Now, I began this post mentioning Apple’s move to a generative AI based technical support system. I find this greatly disappointing. In general, because I believe this type of cost cutting move represents a decline in how I value a company because it represents a decline in how these companies view their customers. And in Apple’s case, I have typically found Apple’s tech support to be better than most companies. Not perfect. Still flawed. But better than most.
I’ve had to work through some tricky issues over the years with Apple’s technical support. Some easily resolved. Some not so. One of those issues required several months worth of conversations, eventually reaching up the chain to Craig Federeghi, before getting the issue resovled. I can’t imagine any tech support system being able to measure up to what Apple has currently, even with flaws and failures that require steadfast persistence and often negotiation.
In my attempts to aid my clients I’ve tried some of the AI chatbots that aren’t part of the specific company I’m working with. Typically they spit out much of what I imagine is in the documentation most phone bank customer service reps use to begin a diagnostic process. (Yes, the device or appliance is plugged in.)
I imagine that Apple will use this new system to handle those rudimentary requests that come in that I’m sure fill up the tech support queues and can be resolved easily. I can only hope Apple then passes more difficult issues to an actual human being who can do the type of intervention necessary to reach a resolution. Heaven help us all, and Apple, if they ever automate the degree of support that requires screen sharing over to a robot.
I have yet to meet anyone in the real world that thinks AI on any level is going to fulfill anything close to the promises made by those pushing it. Folks have already seen and felt the results of these early efforts and have turned sour on the entire concept. I think of this like I imagine the last century’s transition to the automobile. Regardless of how some felt upon seeing the first automobiles on the same streets as pedestrian and horse traffic, everyone knew, for better or worse, it was a future they would have to adapt to.
I have encountered a few business owners who think AI will help them create efficiency and reduce costs, but their discussions about what is possible demonstrate a real hype-fed ignorance and hoped for way of cutting costs. I have one client who is smitten with the promise, but is constantly getting bitten by the results.
I’m betting Apple’s move will take a bite out of its customer satisfacton numbers as well.
(Image from Munthia on Shutterstock)
You can also find more of my writings on a variety of topics on Medium at this link, including in the publications Ellemeno and Rome. I can also be found on social media under my name as above.
Life on the Wicked Stage: Act 3 