Zach McGeary

I've spent 25 years on both sides of post-sales. As the lead customer service and support analyst at JupiterResearch, later part of Forrester, I studied how companies serve customers, back when the question was how many calls a website could deflect. As a strategist at Digitas, I helped them design it. As head of global support and services operations at Cornerstone OnDemand, I ran it, for a 150-person team at a company that grew from eight to ten figures in revenue. Since then I've advised SaaS support teams on the systems underneath. Most recently, I was Vice President of Support at First Due, a public-safety software company.

Outside software, I've built and run operating businesses, scaling companies to eight-figure revenue five times, so I read cost to serve the way an owner does.

Why this practice

Companies are cutting the people who run support before they've built the team that runs the AI. Most AI business cases count tickets: AI takes 20% of them, so cut 20% of the team. But a ticket isn't a unit of work, and AI creates new work that needs owners.

The decisions in front of post-sales leaders now are too big to make on ticket counts. This practice measures the work, prices it, and builds a case your finance team can check. The full argument is in the manifesto, and I publish the research as I go in Proof of Work.

How I work

• Readings, not adjectives: every number shows its sample, window and accuracy.
• Your numbers stay yours: dollar inputs come from your finance team, not from me.
• I'll tell you when a module isn't worth buying.