The tool stopped working for one customer. Just one.
Support checked the usual thing first: was this a known issue. It wasn’t. No other account was seeing it. No spike on the status page. No open incident. Whatever this was, it lived entirely inside one customer’s account, and nobody on the support side had seen it before.
That’s usually where a ticket sits for a while. Not urgent enough to page anyone. Not common enough to have a playbook. Just strange enough that support moves to the next case and hopes it resolves itself.
It didn’t resolve itself. The customer called their AE. The AE called me.
Here’s the part that stings.
Nobody was wrong. Support triaged it correctly. Engineering had no alert to act on because nothing was actually down. Sales did their job and escalated to the person the customer trusted. I did what a CSA does and picked up. Every individual response was reasonable. And none of it got the customer an answer.
Because the answer wasn’t in any one system. It was in four of them.
Support had the ticket, but not the customer’s actual workflow. Engineering had the logs, but no idea what the customer was trying to do when the stall happened. Sales had the context on why the customer bought the tool in the first place and what they’d been promised it would do. I had the relationship, and increasingly, the job of holding all three of those pieces in my head at once.
So I did what you do when no single system has the answer. I built a room. One Slack channel, engineering, support, sales, and me, and I spent the first twenty minutes just getting everyone to describe what they each already knew, out loud, in the same place, for the first time.
It turned out the stall only happened because of how this specific customer had configured the tool, doing something the product technically allowed but nobody had built or tested for at that scale. Not a bug in the way any of our systems would have flagged. A gap between what the product assumed and what this one customer actually needed it to do.
Once everyone’s piece was in the same room, the engineer found it in under an hour.
The hour before that took most of a day.
That’s the part that never shows up anywhere. Not in the postmortem, which focuses on the fix. Not in the ticket, which closes as resolved. The day I spent finding the right four people and getting them to describe the same problem from four different vantage points doesn’t get logged as anything. It just happens, quietly, and then it’s someone’s job to catch up on the rest of the week.
That’s the scavenger hunt. Not a bug hunt. A hunt for whoever already knows the piece you’re missing, run separately, from scratch, every time this happens.
And this happens more than the postmortems suggest. Coveo’s 2025 relevance survey found employees losing an average of three hours a day just searching for information they need to do their job. Microsoft’s 2025 Work Trend Index found the most interrupted employees getting pulled away roughly every two minutes during core hours. The attention research out of UC Irvine puts the recovery cost at 25 minutes to fully return to what you were doing before an interruption like that hits. I didn’t get interrupted once that day. I got interrupted continuously, by my own hunt, for hours.
Forrester’s research on escalations backs up what that day felt like. Only about half of escalations resolve on the first escalated contact. The average takes almost three. Every one of those contacts is someone starting the search over, because the last person didn’t have the full picture either.
None of that is a training problem. It’s not a communication problem. It’s four systems that each hold a true piece of the story and none of them holding the whole thing. The CSA is usually the one who ends up stitching it together, because we’re the one person who’s supposed to know the customer well enough to notice when nobody else does.
Gainsight’s research found that 59% of CS leaders now say scale and efficiency is their top priority for the function. I don’t think that’s a coincidence. It’s an entire industry noticing the same thing at once: the job outgrew the tools a while ago, and somebody still has to hold it together by hand.
If your CS team is building a Slack channel from scratch every time this happens, and losing most of a day before anyone even agrees on what the problem is, that’s exactly who we’re thinking about.