UNRESOLVED
Proactively spotting the moment a customer experiences an issue
Sarala Conlan’s six-person team at Kojo closes 1,300 tickets a month, automates a third of them with Pluno, and pushes database fixes in 20 minutes instead of two weeks. She walks through the stack that gets her there, and the two problems she hasn’t solved yet.
When the customer experiences the issue, you don’t see it
Sarala’s stack tells her a lot. Zendesk holds the ticket. Full Story holds the session replay. DataDog holds the engagement data. Pluno deflects 30 to 35 percent of tickets without anyone touching them. But none of that answers the question she keeps coming back to: at what exact moment did the customer hit the wall?
Right now, finding that moment means a Specialist opens the ticket, reads the customer’s description, then goes hunting through a session replay to match the words to the clicks. It works. It’s also slow, and it happens after the customer has already typed out their frustration. Sarala wants the breakdown point surfaced to the Specialist before the back-and-forth starts. The clicks before the rage click. The screen the customer was actually looking at when something went wrong.
She thinks the answer lives somewhere between AI ticket analysis and session replay tagging. She hasn’t found a tool that does it yet.
Can your stack tell you why a customer is upset before your agent even opens the ticket?
What Sarala’s team actually built
Sarala joined Kojo as the only person in support. She’s now leading a team of six handling 1,300 tickets a month, more than double their late-2024 volume, while holding CSAT steady. The way she talks about that growth says more about her hiring philosophy than her org chart.
She doesn’t think about team size as a headcount question. She thinks about it as a capability question. What’s each person on this team capable of building, fixing, or teaching themselves? A few examples she shared:
- One Specialist earned a Harvard SQL certification on his own time, during PTO, to level up faster.
- Another built a ChatGPT-powered workaround for document scanning issues to unblock customers while engineering worked the root cause.
- A third created the database-change workflow that turned two-week waits into 20-minute fixes.
The tooling reflects the same instinct. Zendesk for tickets, after migrating off Front and Intercom in late 2024 for better analytics. Full Story and DataDog for observability. Jira for engineering escalations, which is about 15 percent of tickets, mostly ERP integration work. Claude and Unblocked for internal product knowledge. Pluno running deflection on the front end and acting as a co-pilot for live Specialists, including auto-creating Jira tickets and reopening Zendesk threads when engineering pushes an update.
Sarala’s view on the stack: the ticketing system is table stakes. The interesting question is what your stack can tell you about why a customer is upset.
On churn, she’s working the same proactive angle. Ticket pattern analysis flags accounts filing repeated high-priority issues, especially around ERP integrations, which is Kojo’s number-one contact reason. NPS detractor scores trigger an investigation workflow that pulls session replays and usage data automatically. At-risk accounts get dedicated Slack channels that loop in support, CS, sales leadership, and sometimes product or engineering, so everyone’s looking at the same picture.
She’s honest about the limits. One account churned despite the signals being there, because by the time the team surfaced the full history — a year of integration timeouts, a champion who left, mounting product frustration — the decision was already made. That case is part of why she’s pushing for earlier signal, not just better triage.
Sarala’s two unresolved problems are the same problem in different clothes. She wants to see the moment a customer breaks down inside the product, before the ticket. She wants to see the personality fit of a candidate, before the hire. Both are about catching signal earlier in a process that currently waits for friction to surface itself.
If you’ve built either of those workflows, or you’re closer to building them than she is, she wants to hear from you. That’s what Unresolved is for.
Links mentioned
Follow the show
Have a story to tell?
So many CX leaders are testing with AI and rebuilding with AI as their foundation. Every one of us have to reinvent everything that we’ve known about customer experience systems and processes.
You’re not alone, so let’s do this journey together.
Be a guest