A Claude Skill that generates a renewal playbook for any at-risk account: which executive to engage, what value story to lead with, which adoption gaps to address before the renewal call, and a 30/60/90-day intervention plan. Takes the account context and turns it into a CSM-ready action sheet.

What you’ll need

Gainsight with health scores, risk flags, and success-plan history
Gong access to the last 180 days of calls on the account
Claude Code or Claude.ai with the Skill installed
A library of renewal playbooks per risk archetype (champion lost, low adoption, pricing pushback, competitive)
Renewal date and ARR per account

Setup

Build the archetype library. Edit references/renewal-playbooks/ with one Markdown file per archetype. Each describes the typical pattern, the recommended interventions, and a sample messaging arc.
Install the Skill. Drop renewal-playbook.skill into ~/.claude/skills/. Set GAINSIGHT_TOKEN and GONG_API_KEY.
Run on a flagged account. build_renewal_playbook(account_id="...", renewal_date="2026-08-15"). The Skill diagnoses the risk archetype, then generates a tailored playbook.
CSM ownership. The playbook is a draft; the CSM customizes and executes. Track interventions as Gainsight CTAs.

How it works

The Skill first diagnoses. It pulls health-score history, success-plan progress, recent calls, and support-ticket patterns, then runs a Claude classification pass to assign one or two archetypes. A “champion lost” archetype looks like a sponsor LinkedIn departure plus a drop in stakeholder calls. “Low adoption” is a usage trend below baseline despite onboarding completion. Etc.

Once the archetype is assigned, the Skill pulls the matching playbook template and tailors it to the specific account: real names of stakeholders, the actual usage gaps, the specific competitor mentioned in calls (if any). Each intervention has a date, an owner, and a measurable outcome.

The 30/60/90 view forces structured pacing. CSMs tend to either over-react in week one and burn the relationship, or do nothing until 30 days before renewal. Both fail. The plan paces it.

Watch-outs

Archetype overlap. Many at-risk accounts have multiple causes. The Skill returns up to two; force-pick one as primary or the playbook becomes mush.
Plan vs. execution. The playbook is worthless if interventions aren’t tracked and reviewed. Tie it to Gainsight CTAs.
Renewal-date math. Don’t generate the playbook 30 days before renewal. Run at 90 days minimum to leave room for actual intervention.

Stack

Gainsight — health, risk, success-plan history
Gong — call evidence for archetype diagnosis
Claude — diagnosis and playbook tailoring