How to use the Chief Customer Officer Advisor agent skill

What the skill makes the agent do

The skill covers four decisions: how your retention is built and whether gross retention and net retention are telling the truth, how to segment customers for differential investment, which coverage model your CS team should use (pooled or named account managers), and which customer success role to hire next.

When you ask about one of these topics, the agent works through a fixed framework. It checks gross retention before net retention, splits churn into root-cause categories, sorts customers into investment tiers with an ICP fit score, and compares pooled versus named coverage. It can also run the included Python scripts on sample data or your own JSON files so it produces numbers instead of guesses.

• Retention decomposition: gross retention, contraction, expansion and a churn root-cause taxonomy
• Customer segmentation: tier design, ICP fit scoring and investment per segment
• CS team coverage model: ratio math and thresholds for pooled versus named CSMs
• CS team org evolution: how customer success differs from Support, Account Management and Implementation

Install the skill from the ZIP download

Download the ZIP of the skill from this page and unzip it. To make it available to a single project, place the unzipped folder inside that project's .claude/skills/ directory. To make it available across your projects, place it in ~/.claude/skills/ instead.

If you use the Claude apps rather than a project folder, you can upload the skill under Settings, then Capabilities, then Skills.

• Project only: put the folder in the project's .claude/skills/ directory
• All projects: put the folder in ~/.claude/skills/
• Claude apps: upload the skill under Settings → Capabilities → Skills

Trigger the skill with natural requests

You do not need to name the skill. The skill description says to use it when designing retention strategy, segmenting customers for differential investment, sizing the CS team, or sequencing CS hires. Describe your situation in plain terms and include your numbers where you have them.

The skill's key questions are a good guide to what to provide. Start with your gross retention rate, your main reason for customer churn, your time-to-value by segment and your ARR-per-CSM ratio.

• "Our net revenue retention is 115%. Is our retention actually healthy?"
• "Help me split our customers into tiers and decide how much CS time each tier should get."
• "We have 350 customers and a small CS team. Should we move to pooled coverage or named CSMs?"
• "We have one CS person already. Should our next hire be a CSM, a support engineer or an account manager?"

Run the included scripts on sample or your own data

The skill includes three Python scripts. Each one runs with no arguments on an embedded sample, or with the path to your own JSON file. The scripts are documented as using only the Python 3 standard library. Your JSON files should follow the schemas described in the reference files.

Retention decomposition uses cohort data. Segmentation design uses customer data. Coverage calculation uses book data describing your customers.

• Retention: python scripts/retention_decomposition_analyzer.py (sample) or python scripts/retention_decomposition_analyzer.py path/to/cohorts.json
• Segmentation: python scripts/customer_segmentation_designer.py (sample) or python scripts/customer_segmentation_designer.py path/to/customers.json
• Coverage: python scripts/cs_coverage_calculator.py (sample) or python scripts/cs_coverage_calculator.py path/to/book.json

Check safety, scope and what the skill does not cover

The skill's manifest says it runs scripts and does not use the network. Read the scripts before you run them on company data, and keep your JSON inputs to the documented shapes. Treat the output as a structured starting point, not a verdict.

The skill is strategic only. It does not cover tactical CS implementation such as health-score tooling, CRM workflows, NPS survey setup or onboarding automation. The skill points to a separate customer success management skill for those tasks.

• Scripts run locally and the skill does not use the network, according to its manifest
• The thresholds are described as generic B2B SaaS startup baselines, so validate them against your own data
• Do not make firing or headcount decisions from sample outputs alone
• Large enterprises may need different benchmarks

Tips for getting useful output

Give the agent real numbers from your own data rather than estimates, and say which segment or cohort each number describes. Ask for gross retention before net retention so the agent reviews churn and contraction first. When the answer depends on stage and deal size, say so, because the skill notes that both determine whether pooled or named coverage fits.

• Provide gross retention, churn reasons and time-to-value by segment
• Ask the agent to run the matching script on your JSON file, then explain the result
• Ask which assumptions in the output you should validate first