‍The Resolution Gap: What Customers Want vs. What Vendors Measure

Your card gets declined at the pharmacy counter, and your kid needs an urgent prescription. Rent is due tomorrow, and the paycheck hasn't landed. Your car dies on the shoulder of the highway at 11 pm. In all three scenarios, there is urgency, and if the voice on the other end doesn't show understanding and instead insists on reconfirming why you're calling, something inside you tells you this isn't going to end how you want. 

The customer service industry has spent 20 years refining and optimizing for getting people off the phone instead of solving their problem. 

• It started with the IVR: press 1 for billing, press 2 for support, press 0 to talk to a person. That last option has been an escape hatch, an admission the menu can’t solve everything.
• Then AI arrived. NLP, ML, and conversational interfaces. The pitch changed. The product didn't. A new generation of vendors started competing on containment rate, the percentage of calls kept away from a human, the same finish line the IVR was chasing, just with better manners.

This is how the market is wired.

Start with the buyer. 91% of customer service leaders report pressure from executives to implement AI, and over 80% of organizations expect to reduce agent headcount within 18 months. When the business case is written in headcount and cost per contact, the metric that wins is the one that maps to it: how many contacts never reached a person. In contrast, Gartner's consumer study found that only 14% of issues are fully resolved in self-service, and customers have to either self-serve or find a solution across multiple channels. That is the containment gap in one statistic. The interaction gets counted as handled, the problem isn't solved, and the customer is back.

The Resolution Gap

At Replicant, we wanted to know if that gap was real or just a hunch, so we surveyed over 1,000 U.S. consumers on what they actually want from customer service AI. View the full report here.

• What "good" means to a customer:
  
  • 69% — my problem got solved
  • 13% — politeness
  • 12% — fast
  • 6% — didn't have to repeat myself
• The number that should worry anyone selling containment as the finish line: 
  
  • 90% of consumers already assume, before they've said a word to an automated system, that they'll need a human anyway.
• And when the bot can't resolve something: 
  
  • 27% say escalating to a human is their #1 frustration with automated phone systems
• The line that reframes everything: 
  
  • 57% don't care if it's a human or an AI, as long as the issue gets fixed.

Line those numbers up against the metric the industry actually reports, and the mismatch is the whole story. 69% of customers define success as solved. The industry defines success as contained. Those aren't two different things, and only one of them is what the customer asked for. 

Containment won as the industry standard not because it's what people want, but because it's easier to measure: did the call end without a transfer, yes or no. Resolution on the other side asks a harder question: did the issues actually get fixed?. Answering it means building something that reaches into the backend, takes the action, and confirms it worked. That is why a lot of vendors in the industry chose the easy number (containment) instead of the one users want (resolution).

That's not entirely unreasonable. Some resistance to measuring resolution is fair; a fraud dispute or a hardship exception needs human judgment, not a bot making the decision. But that's a small slice of the contact volume. A password reset, a billing correction, an appointment move, an eligibility check, these are high-volume, rule-based workflows. Resolution isn't hard to measure. It's just harder than counting whether a call got transferred, and "harder" is exactly why the industry has spent two decades not doing it.

This is the gap Replicant was built to close; not the exceptions that genuinely need a human's judgment, but everything else: the repetitive, high-volume requests that make up most contact volume. We resolve those end-to-end and confirm it against the system of record. The result is for people to spend their time on conversations that actually need a person, not redoing the work automation already touched and failed to finish.

Back to the Call That Mattered

Go back to the pharmacy counter, the rent due tomorrow, or the car on the shoulder at 11 pm. Those people were asking for a solution, so they could stop thinking about the problem. And if you look at those three moments again, something signaled the problem before the interaction started. The card on file was expiring. The deposit was running late. The car was showing warning signs. A containment mindset never sees any of that, because it only starts counting when the conversation begins. A resolution mindset starts from the problem.

That's the shift vendors need to make. Solutions exist to solve problems, and the best ones solve them before the person has to ask. Vendors who can resolve end-to-end workflows are the ones that solve with a proactive service mindset.

Resolution should be the metric every conversational AI and CX vendor is judged on because it is the one customers care about. Did we solve the problem, and could we have solved it before they had to contact us? The best customer service interaction was never the fast one. It's the one that never has to happen a second time, and when we do our job well, the one that never has to happen at all.

‍