← Results
Two-site national operation · 5-person customer service team · Leadership role

The team that was shrinking while the work grew

A shrinking customer service team facing rising ticket volume — instead of fighting for headcount, we studied the work, automated the repetitive parts, and had the team help build its own tools.

The situation

What we started with

A small customer service organization moved under my operations umbrella at exactly the wrong moment: ticket volume was climbing, the work was getting more complex, and attrition had already shrunk the team — with no backfills coming. The default move was obvious: go fight for headcount.

What I found

On the floor and in the data

Before asking for people, I studied where the existing hours actually went. I did portions of the work myself to learn it firsthand — and the pattern was hard to miss. A large share of the team's day was repetitive, manual, and highly structured: transferring tickets between systems that didn't talk to each other, and typing out near-identical responses to the same customer situations. This wasn't judgment work. It was routing and repetition wearing a judgment-work costume.

What we changed

The work

I partnered with our engineering team and dug into agentic AI workflows — not by buying a tool off the shelf, but by automating the specific, structured tasks the diagnostic had surfaced. The overlooked step: I rebalanced the team's labor to create capacity for people to participate in building the automation, so the team shaped its own tools instead of having them dropped on their desks.

The results
~50%
Nearly 50% of the team's manual labor hours automated across its core functions
0
Zero headcount added while volume and complexity kept growing
+scope
The recovered capacity let the team absorb additional functions beyond its original scope
2 hires
On a five-person team, that's not an efficiency stat — that's two hires that never had to happen

The part that applies to your building: when a team is underwater, the first question isn't "how many people do we need?" — it's "where do the hours actually go?" Most operations answer with a hiring req or a software purchase before anyone's studied the work. The study comes first. Sometimes the answer is people. Sometimes it's automation. It's never a guess — and that's fixable too.

Something going sideways in your operation?

The first call is a real conversation.

Book a first conversation