A.I. is coming for Customer Service jobs

A.I. is replacing customer service people.

While attending trade school I encountered the Commodore Vic-20 computer. The school used it to teach about using computers and basic programming languages. I had to have one, all 5k of RAM. This was the year 1980. I would type in free game code, listed in geeky magazines, there was no home Internet yet. I read every computer magazine I could get my hands on to learn the latest news. The listings were almost always in the BASIC programming language. BASIC was an acronym for Beginner’s All-purpose Symbolic Instruction Code. This came in handy later supporting IBM PCs and their clones, whose operating system was MS-DOS. I learned to write complex batch file programs in DOS. It’s pronounced daahs. Think of when a doctor checks your throat and you say, “ahhh”, but don’t hold the sound that long. MS-DOS stands for Microsoft Disk Operating System.  Programming DOS was similar to BASIC.

Having graduated tech school, I went into a computer retail chain store known as ComputerLand to look for a job and met one of the owners. While talking he told me he had a degree in electronic engineering, and was a tech at heart. He hired me to work in the stock room. I helped the service department when pre-Christmas sales made us very busy. We prepared and configured computers according to individual sales specifications and made deliveries which included setting up the computers and printers for customers. Later the service manager lobbied to make me a full-time tech. ComputerLand was an authorized IBM PC and Apple computer sales and repair center. The IBM PC was only a year old. This was 1983 and the Mac was still an Apple secret. 

Apple offered schools a discount on the Apple IIe and would ship the orders directly to the schools and notified us of deliveries. We would then make an appointment to go to the school, unbox and set-up their new Apples. Apple was smart to give them a discount and get Apple product into the hands of the next generation. They also offered a discount to teachers to buy a computer of their own for home.

Apple did sell a computer with a similar graphic interface as the Mac, before the Mac was introduced. It was called The Lisa. It was much larger, heavier and more expensive than the Mac. The Lisa was a Big Mac.

Steve Jobs wanted Microsoft business software available on the Mac. This would increase the Mac’s credibility in the business world and present the Mac as more than a cool, very expensive toy for home. 

Bill Gates wanted to to gain insight into the Mac’s WYSIWYG (What you see is what you get) operating system and develop the graphics based Excel spreadsheet. I’m guessing Apple programmers were instructed by Jobs to cooperate with Microsoft programmers. This would help Microsoft in developing the MS-Windows operating system. I have no proof of any of this, but it seems likely. 

I don’t remember applying for a job at Apple, maybe I applied online. When Apple called me, I was both excited and panicked. I had been working contractor jobs, supporting PCs and Windows in corporate environments for years now. I had lost touch with all things Mac. They offered Skype or FaceTime for the interview, but the only Mac I had didn’t run FaceTime, not by a long shot.

The job was for the position of AppleCare At-Home-Advisor.

When I met my fellow trainees, they were PC and Android people for the most part. Apple provides training which starts at square one and they offer a very generous employee purchase plan after an initial period of time. Shortly before training begins, they ship you an iMac for work and an iPod touch and headset for your landline phone. Apple later migrated us to VOIP so we could plug a new headset into the iMac and not need the landline. The iPod touch does everything an iPhone does, except make cell phone calls. It can make FaceTime calls if you have Wifi available, so having one was great for iOS support. Eventually I did buy an iPad Retina and a MacBook Pro with my employee discount. It took at over a month before I finished interviewing and started training. The day for the Skype call came during the two week period I had prepared a show for the interviewer. I borrowed a few Apple products, which needed to be returned to stores within two weeks, so that before the call, I had the stage set. I framed the Apple products I had in the background. On a desk behind me stood my original Mac, a MacBook Pro, an iPad, an iPod Touch, all with colorful, swirling screen savers or home screens lit up. There’s no business like Show Business! The Apple interviewer was no stranger to stage craft either, he had various engraved glass trophies from Apple, in camera view, on his desk. Lucky for me, we had good chemistry and I was somewhat relaxed despite the circumstances. 

He asked me a couple ‘what if’’ troubleshooting scenarios which were familiar to me. They were general questions that could apply to any kind of computer, Mac or PC. His first question was “What would I do if the customer’s computer wouldn’t boot?” Not knowing enough about OSX recovery mode, safe mode, or PRAM and SMC resets, I said, “I’d try turning off the Mac then unplug all USB devices, then try to start it up. That seemed to go well, then he wanted to role play. He said he’d play the customer, calling me for help. “Ring”, “Thank you for calling Apple, how may I help you?”, I said pretending to be a customer, he said something like, “I lost my iPhone, what can I do?” I remember the first thing I said was, “I’m really sorry to hear that.” – He immediately gave me a thumbs up. Empathy, I later learned, is Job One with Apple. The dogs and cats were banned from my room during the interview and no one barked.

A few days later, I received another call and was hired, training started in a couple weeks. I continued to practice with the Apple products and dived into every question that arose while using these amazing gadgets. I went Google nuts, and used far too much ink, printing information and organizing small notebooks into every topic that came to mind. These habits continued through training: notes, flowcharts and notebooks with tabbed categories. I didn’t need to do that as Apple has a rich, deep, searchable online knowledge base for advisors and lots of troubleshooting flowcharts to follow. The bad side is that the knowledge base will soon be used to replace level one advisors, as soon as artificial intelligence can fool people into thinking there is an empathetic, human on the other end of their call. 

Customer support is all about If and Then flowcharts. If the problem is A, then try B and if Y happens, try X or if G happens, try Z. This is the same as programming. A.I. will be as happy as a pig in mud doing customer support. A.I. is very good at this sort of thing.

I also realized that by filling out forms for each of my calls, I was creating notes that A.I. could follow as were all my co-workers. This might help explain when filling out call reports was so stringent and exacting.

Everyone worked at least one weekend day. Everyone bid on a new shift every six months. 

People complained and had every excuse. 

These won’t be problems for A.I.

When a customer called AppleCare, the call would be routed to the next available advisor, if I was in ‘Available’ mode and I was next in line to receive a call, I would answer the call, “Thank you for calling Apple, my name’s John, how can I help?

I’m sure A.I. will offer the same greeting with some bogus name.

A Vendor is a big company that has a contract with Apple to hire and train people to do the exact same job that Apple employees do, with the exact same software and connections to Apple and the callers call the same number; 1-800-MYAPPLE. When someone calls Apple, they often get a vendor employee who says, “Thank you for calling Apple.” The only difference for the employee is about half the pay he would get from Apple and they offer anemic, if any benefits and no employee purchase plan. Apple pays the vendor more for you than the vendor pays you. It saves Apple money as well as the cost of benefits. I think they will get what they pay for, a lack of loyal, long term, experienced advisors resulting in decreasing public opinion of Apple’s support, but A.I. is on the way to save the day! Save the day for the company management, not the customer service reps. A.I. doesn’t get emotional about wages. I’m using Apple here as an example, because I worked there. All large providers of customer service are looking to A.I. as the next wave.

I logged each call with the customer info, the hardware, the problem, the troubleshooting steps I took and the solution. The computer form had to be filled in completely. It would not save and I could not go back to ‘Available’ status if I skipped filling in the form. Apple tracked number of calls verses number of completed forms as a metric also. Apple has excellent software, albeit often obstinate and stubborn. Often as I filled the form out, it would pop-up more questions with blanks to fill in. Often questions that seemed endless. It insisted on having everything answered which often screwed up my ACW metric (time between calls) and slowed things down. No one liked the logging software. I was expected to type the form in as I spoke to the customer, before I understood the problem or had the solution. I don’t chew gum and walk well at the same time while troubleshooting. When I’m listening to a customer’s issue, I’m listening and looking for clues to help me probe deeper into the problem and get to a solution. I’m there for the customer, not the damn metrics! So I often would fix the problem, let the customer go on his or her way and then I finished filling out the form after the customer hung up. This causes a longer ACW average as ACW is measured from the time between when the customer hangs up and the time I signal I’m available for the next caller – Available mode. It was even suggested to me that I keep the customer on the phone longer than I needed to. That way I could finish filling in the form before they hung up and thus shorten my ACW average, I kid you not. That’s not customer service to my mind. Two symbiotic metrics were how often I transferred to Tier 2, senior advisors and how long my average call length was. 

CSAT or Customer Satisfaction was also a metric. If I didn’t handle a call just right or the customer was not happy with Apple for any reason – bad CSAT. It could be due to the hold time before the call even reached me, or some other reason, it didn’t matter, there goes your CSAT survey metric.

The logging requirement won’t stop after A.I. handles the calls, it’ll get better and so will the A.I.
The technology is already here. Corporations would prefer it if the general public would accept it, but until then, they’d rather not get caught using it. Most people need to get used to the idea, but they will. The solution is to fool the caller into thinking they are speaking to flesh and blood people. That technology is here now and it understands context and nuance and can imitate speaking like a regular, flawed human. If you don’t believe me, Google: “Google Duplex: A.I. Assistant Calls Local Businesses To Make Appointments”. I’m assuming companies won’t have to pay A.I. by the hour and no healthcare. 

Also see IBM’sWatson doing customer service.

The customer will benefit by longer time allowed for calls and presumably A.I. won’t get stressed, but if it does, look out!

IBM Watson: Final Jeopardy! and the Future of Watson -this was 9 years ago

This entry was posted in Uncategorized. Bookmark the permalink.