The Easy Ones | Customers, Etc.
We naïvely think that taking care of easy problems is going to solve all our problems.
Last Friday, I was at a pre-Advent ugly sweater Christmas party at my neighbor’s house. A few of us started talking about ChatGPT and the recent advancements in AI. My neighbor is a radiologist, so I asked him if he had started seeing AI impact his work.
“Yes”, was his answer, but only for the most obvious issues. It wasn’t helpful in more nuanced cases that required deeper analysis, and even for the simpler instances, it barely saved him any time. If his technicians shared the data with him just slightly differently, he would be faster than the AI. While he recognized that AI could eventually evolve to be more advanced in radiology, right now it’s just focused on the easy problems, and it’s unclear how much value is being added.
I was instantly intrigued by his perspective. In my work in customer support, I can’t tell you how many times someone has asked me how much bandwidth would be freed on the support team if we could find a magic solution to take care of “the easy ones.”
“The easy ones”
I get irrationally bothered when an executive or founder zeroes in on “the easy ones” as the thing that’s the most obvious problem for the support team to solve. If we just get rid of the easy ones, we’ll free up so much bandwidth on the team and won’t have to scale the team linearly. You weren’t thinking of scaling the team linearly, were you? Gasp.
The solution for the easy ones have changed over the years: we’ll add a help center; we’ll add community support; we’ll outsource1 ; interns!; we’ll use a crappy chatbot; we’ll use a “ticket-deflection tool” in our help center that makes you fill out an entire email before it tries to be helpful; we’ll use a less crappy chatbot; we’ll improve product quality so the easy ones don’t show up in the inbox (and how!); we’ll use ChatGPT, but for our product.
There’s nothing new under the sun. Let’s explore a few simple misconceptions with the easy ones.
Myth: The easy ones are a major pain
They’re not. The easy questions are the ones we answer when we’re having our coffee in the morning. They’re how we warm up and get ready to tackle the hard problems. (They might even be a welcome excuse to avoid working on the hard problems. Ahem.)
If you open the support inbox in the morning and there are a dozen tickets and eight of those tickets are easy, you’re positively thrilled to fly through those tickets before you have to sink your teeth into the harder problems. Sure, you want to improve the product, add documentation, etc. etc., to reduce those tickets from showing up in the inbox in the first place, but on the whole, they’re not a serious bother.
Myth: Fixing the easy ones makes all your metrics better
Heh.
What do you think is going to happen to your Average Handle Time (AHT) metric when you eliminate the easy issues from the inbox? Basic math demonstrates that AHT will increase if you get rid of the easy issues. This is totally fine—you’re fixing bugs, getting answers to customers faster in a self-serve manner, etc.—but you still have to be aware that not all your metrics will go in the direction you expect.
Myth: Fixing the easy ones opens up untold stores of bandwidth
“What would you do with all of the free time your team has when half the inbox is solved by our fancy new AI tool?”
There’s an unhelpful false equivalence that all support interactions are created equal. This is the myth that’s at the root of all of the myths we’ve been talking about.
Look, I’ll take a bump in efficiency any day of the week, but don’t sell it to me like I’m going to be able to instantly lay off repurpose half my team. For starters, the bandwidth you save because easy tickets aren’t in the inbox will need to be spent on doing quality control of the AI. If you’re not babysitting your chatbots, you’re just transferring your handle time from your agents to your customers.
Reality: Where AI is headed
Do I think AI won’t be useful? Absolutely not. Just because I have a curmudgeonly reaction to naïve assessments of the impact of AI on team productivity doesn’t mean I don’t believe AI will add value. Nor do I believe it won’t have a major impact, in time.
Do we want the easy tickets to go away? Of course we do. But not for the wrong reasons. If we’re doing it right, the easy tickets are going away because customers are having a better experience. They’re getting helpful answers to their questions, faster, or better yet, they’re not having to reach out for help at all because the product does a good job supporting itself and that in turn leads to a low contact rate in the support inbox.
In my post last week, we looked at Scott Hanselman and Mark Downie’s demo about AI. Scott talked about how conversational tools like Microsoft Copilot and ChatGPT are a new kind of user interface. We’re still in the process of figuring out how this new tool will be useful.
As with any tool, when it comes to the context of customer support, the goal isn’t to provide the best possible support tool for your product. The goal isn’t even about providing the best possible product. The goal is help users feel that they’re awesome at whatever it is they need to be doing2. Our tools just need to do their job and get out of the way. The goal was never to get rid of the easy ones.
If you only think outsourcing is for the easy ones, that’s a problem. Let’s talk.