AI in CS
The QBR Transcript Pipeline: Sentiment as a Leading Indicator
A working pipeline that turns recorded QBRs into a sentiment delta — and the four pitfalls that make every off-the-shelf version of this useless inside an enterprise account.
The AI Pillar · A working publication on AI for CS operators
You've seen the AI demos for CS. This is about the AI you can actually put to work.
Frameworks, working notebooks, and teardowns for the AI that actually lands inside a CS function — not the keynote version. Below: the five-signal renewal model, as a diagram and as code.
Model · v0.4 · trained on 312 enterprise accounts
The job of any CS model isn't a number — it's an defensible prediction. Inputs on the left are the signals every CSM already collects. Outputs on the right are the only three the business actually buys.
x1
Product usage depth
feature breadth · DAU/WAU
x2
Sponsor engagement
exec touch · multithread depth
x3
Support velocity
P1 frequency · time-to-close
x4
Sentiment delta
NPS · QBR transcripts
x5
Renewal proximity
days to term · auto-renew flag
ŷ₁
Churn risk
0%
ŷ₂
Expansion fit
0%
ŷ₃
Days to escalate
0d
Inside the AI pillar0 pieces · Updated weekly
AI in CS · Flagship
The Five-Signal Model: A Neural Net Your CRO Will Actually Approve
The diagram above is the entire thing. This piece walks the architecture choice, the training data, the calibration curve, and — most importantly — the conversation you have with the CRO the first time it disagrees with a CSM's gut.
Teardowns
Four CS Copilots, Tested Against a Real Renewal Quarter
A methodology-first scoring framework — six dimensions, weighted, scored 1–5 — applied to four representative copilot categories as worked examples. The scorecard is the contribution.
AI in CS
The AI Decision Matrix: What CSMs Should Automate, Augment, or Leave Alone
AI tooling decisions in CS should be made one task at a time, not one vendor at a time. The 4×4 matrix that tells you what AI should automate, augment, inform, or leave alone.
AI in CS
The CSM's Prompt Library: Eight Prompts That Earn Their Place in the Workflow
Most published prompt libraries for Customer Success are useless because they're written by people who have never run a renewal. Eight working prompts, organized by Decision Matrix quadrant.
AI in CS
AI Safety for Customer-Facing Work: Eight Rules That Keep You Out of Trouble
Almost every published AI safety conversation is happening at the wrong altitude for CS work. Eight operational rules that produce defensible decisions in under thirty seconds.
Issue № 07 · AI · May 2026 · Written for senior CS operators · No sponsors