AI summary
Seven AI prompts for CSMs in SaaS: account health reads, QBR deck drafting, renewal conversation prep, churn-risk save plans, onboarding from sales notes, customer story drafting, and expansion-opportunity memos. Built to scaffold CSM judgment without papering over real risk.
CSM work lives at the intersection of product knowledge, relationship management, and revenue. The seven prompts below take the parts of CSM work that compound when done weekly (health reads, QBR prep, save plans, expansion memos) and structure them so your time goes to the customer conversations themselves. This is the CSM slice of the AI Prompt Library, paired with a connector callout. For sales-side prompts see Best AI Prompts for Sales.
Why do most AI customer-success-AI workflows produce QBRs nobody reads and renewal conversations the customer wins?
The default CSM-AI risk is letting AI confirm optimistic health reads. Optimistic reads die at renewal. The prompts below are built to push back on optimism and force the harder read of risk while there is still time to act.
Use AI to structure the read; you bring the customer context. If you let AI draft anything customer-facing, run it through How to Edit AI Out of Your Writing first. When a prompt becomes a weekly habit, graduate it using the Prompt-to-Workflow Ladder.
What are the seven for customer success prompts?
Prompt 1
Account Health Read
Most CSMs check health scores once a quarter and miss the early signals. This prompt structures the weekly read so account risk surfaces before the renewal email.
Account I am reviewing (de-identified): ACCOUNT TIER and ARR band: [BAND] ADOPTION METRICS (usage, seats activated, key feature usage): [BRIEF] SUPPORT TICKETS in last 90 days: [VOLUME, CATEGORIES] LAST EXEC TOUCH: [WHEN, WITH WHOM, OUTCOME] CHAMPION STATUS (still there, role changed, departed): [BRIEF] RENEWAL DATE: [DATE] EXPANSION ATTEMPTS, prior outcomes: [BRIEF] Produce a health read: 1. THE CURRENT HEALTH score with reasoning, not just a number. 2. THE 3 EARLY SIGNALS in the data (positive or negative). 3. THE CHAMPION RISK: is the relationship single-threaded. 4. THE EXPANSION FIT: where their behavior suggests upsell or where it does not. 5. THE RENEWAL RISK: realistic read with the specific concerning data point. 6. THE NEXT 30-DAY ACTIONS in priority order. 7. THE ESCALATION TRIGGER that would change this read materially. Do not let me confirm what I want to believe. If churn risk is high, name it.
When to use: Weekly cadence on top 20 accounts, monthly on the rest. · Best model: Claude. The discipline about not confirming wishes matters.
Prompt 2
QBR Deck Drafter
Most QBRs drift into product roadmap show-and-tell. This prompt drafts the QBR around the customer’s business outcomes instead.
QBR prep: ACCOUNT (de-identified): [BAND, TENURE] THEIR STATED BUSINESS OUTCOMES at the start of the contract: [LIST] WHAT HAS HAPPENED with those outcomes (data we have): [BRIEF] KEY USAGE PATTERNS that connect to outcomes: [BRIEF] WHAT WE HAVE SHIPPED in the past quarter relevant to their use case: [BRIEF] KNOWN CHALLENGES they are facing: [BRIEF] WHO IS IN THE ROOM and what each person cares about: [LIST] WHAT I WANT to walk out with: [SPECIFIC OUTCOME] Draft the QBR structure: 1. THE OPENING: 2 sentences naming what matters most to them this quarter, not what we want to show. 2. THE OUTCOMES RECAP: their stated goals with the data we have on each. 3. THE WINS: 2-3 specific moments where the product or partnership produced value. 4. THE GAPS: where we have not yet delivered what was promised. Direct, not buried. 5. THE OPPORTUNITIES: 1-2 paths forward, tied to their priorities. 6. THE ASKS: what we need from them to execute the opportunities. 7. WHAT I WANT THEM TO DO: the specific next step. Do not over-present product roadmap. Customers care about their outcomes, not our launches.
When to use: 10 business days before the QBR. · Best model: Claude. Discipline about not over-presenting roadmap matters.
Prompt 3
Renewal Conversation Prep
Renewal conversations go badly when the CSM enters without a clear picture. This prompt structures the prep so the conversation is grounded.
Renewal conversation prep (de-identified): ACCOUNT: [BRIEF] RENEWAL VALUE: [APPROXIMATE] CURRENT CONTRACT terms: [BRIEF] WHAT WE ARE PROPOSING: [BRIEF] WHAT THEIR CHAMPION HAS SIGNALED about renewal: [BRIEF] WHAT WE HAVE DELIVERED in the past year: [SPECIFIC] KNOWN OBJECTIONS or competing options: [BRIEF] WHAT I CAN MOVE on (price, terms, packaging): [BRIEF] MY WALK-AWAY: [INTERNAL ONLY] Produce a prep brief: 1. THE FACTS: what we have delivered, in their language. 2. THE LIKELY OBJECTION and how to address it without retreating. 3. THE PRICE ANCHOR: how to frame the value before naming the number. 4. THE NEGOTIATION SCRIPT: my opening, my fallback, my walk-away. 5. THE EXPANSION ANGLE: if renewal is locked, where the upsell fits naturally. 6. THE RED FLAGS to watch for in the conversation. 7. THE FOLLOW-UP within 24 hours of the call. Do not let me anchor to optimism. Be direct about whether the deal is at risk.
When to use: 1 week before the renewal conversation. · Best model: Claude. Discipline about realistic anchoring matters.
Prompt 4
Churn-Risk Save Plan
Once an account is flagged at risk, most CSMs scramble without a structured plan. This prompt builds the plan in 20 minutes.
At-risk account (de-identified): WHY IT IS AT RISK: [SPECIFIC SIGNAL: champion departure, low adoption, exec frustration, competing solution, budget cut] VALUE AT RISK: [ARR] TIME TO RENEWAL or contract event: [BRIEF] WHO ON THE CUSTOMER SIDE knows we are paying attention: [BRIEF] WHO INTERNALLY needs to be involved (product, exec, sales engineer): [LIST] WHAT WE HAVE TRIED already: [BRIEF] Draft a save plan: 1. THE ROOT-CAUSE READ: what is actually driving the risk, beyond the surface signal. 2. THE STAKEHOLDER PLAY: who we need to engage on the customer side, in priority order. 3. THE INTERNAL ESCALATION: who internally has to be in the room and by when. 4. THE NEXT-7-DAY ACTIONS: what happens this week. 5. THE COMMITMENTS we can credibly make to the customer. 6. THE GO / NO-GO checkpoint: in 14 days, what would tell us this save is working. 7. THE WALK-AWAY: at what point we accept the loss to focus elsewhere. Do not promise what we cannot deliver. Saves built on false promises return as churn 6 months later.
When to use: Within 48 hours of the risk being flagged. · Best model: Claude. The discipline about not over-promising matters.
Prompt 5
Onboarding Plan from Sales Notes
Sales-to-CS handoff loses context every time. This prompt extracts the actual onboarding plan from sales notes so the customer feels continuity.
Sales handoff notes (de-identified): WHAT THEY BOUGHT: [PRODUCT / SKU] WHY THEY BOUGHT IT (the use case sales described): [SPECIFIC] KEY CONTACTS named in sales process: [ROLES] COMPETING SOLUTIONS they considered: [LIST] TIMELINE EXPECTATIONS sales set: [WHEN THEY EXPECT VALUE] KNOWN CONCERNS they raised: [LIST] WHAT WAS PROMISED in the sales process: [LIST] SPECIAL TERMS or commitments: [BRIEF] Produce an onboarding plan: 1. THE FIRST 30 DAYS: what success looks like, tied to their stated use case. 2. THE FIRST QUICK WIN: a specific outcome they can see by week 2. 3. THE STAKEHOLDER MAP: who we engage with for what, in what sequence. 4. THE PROMISES TO HONOR: explicit list of what sales committed to. 5. THE EXPECTATION RECALIBRATION needed: anything sales over-promised that I need to manage now. 6. THE 90-DAY GOAL: what we want true by day 90 to set up the year well. 7. THE FIRST KICKOFF MEETING agenda. Do not assume sales captured everything. Flag what I should ask the AE or the customer to clarify before kickoff.
When to use: Day of handoff, before customer kickoff. · Best model: Claude or ChatGPT.
Prompt 6
Customer Story Drafter
Marketing wants customer stories. Most CSMs hate writing them. This prompt drafts the story from your knowledge so marketing can edit instead of interview.
Customer story I am drafting (with customer permission, de-identified at this stage): CUSTOMER: [INDUSTRY, COMPANY SIZE] BEFORE STATE: [WHAT THEIR SITUATION WAS] WHY THEY CHOSE US: [SPECIFIC REASON] WHAT THEY DEPLOYED: [USE CASE] WHAT CHANGED for them, with specific outcome: [DATA] KEY MOMENT in the deployment that mattered: [SPECIFIC] WHAT THEY WOULD SAY if I asked them: [QUOTE OR PARAPHRASE] Draft a 400-word customer story: 1. THE OPENING: a specific moment or detail that grounds the story. 2. THE CHALLENGE: what their business needed. 3. THE CHOICE: why they picked us, anchored to specific differentiation. 4. THE DEPLOYMENT: how it happened, with one concrete detail. 5. THE OUTCOME: specific data points and a representative quote. 6. THE LESSON: what this proves we can do for similar customers. Do not exaggerate outcomes. Marketing will tighten the copy; the data must hold up to fact-check. Avoid: "streamline," "transform," "empower," "unlock," "holistic," "end-to-end." Use specific numbers and verbs.
When to use: After customer permission, before marketing interview. · Best model: Claude. The discipline about not exaggerating matters.
Prompt 7
Expansion-Opportunity Memo
Expansion opportunities die in CSM’s heads. This prompt structures the memo so the AE, the manager, and the account team see the same picture.
Expansion opportunity (de-identified): ACCOUNT: [BAND] CURRENT FOOTPRINT: [WHAT THEY HAVE] EXPANSION TYPE: [SEATS / MODULE / TIER UPGRADE / NEW USE CASE] WHY NOW (the trigger event): [SPECIFIC] WHO ON THE CUSTOMER SIDE is the buyer: [ROLE] WHAT WE KNOW about their budget cycle: [BRIEF] COMPETING OR ADJACENT TOOLS they have: [LIST] MY HYPOTHESIS on deal size: [APPROXIMATE] Draft an expansion memo: 1. THE TRIGGER: what makes this expansion timely. 2. THE FIT: why our solution is the right answer to the trigger. 3. THE BUYER: who decides, what they care about. 4. THE COMPETITION: how they would evaluate alternatives. 5. THE ASK: deal size, timing, terms. 6. THE INTERNAL ROLES: AE involvement, SE involvement, exec sponsor if needed. 7. THE NEXT 14-DAY PLAN. Do not anchor to my optimism. Be direct about close probability.
When to use: When the expansion signal first appears, not after the quarter closes. · Best model: Claude. Realistic anchoring matters.
These work across Claude, ChatGPT, Gemini, and Grok. Claude is the strongest default because of its discipline about not confirming optimism. Grok is sharpest for the at-risk save plan. ChatGPT is broadest. For customer data, use paid tiers with no-training and no-retention terms.
What is the worst thing you can do with AI for customer success?
Three patterns will burn CSMs fastest.
- Letting AI confirm an optimistic health read on an account where the data says otherwise. The Account Health Read prompt is built to surface risk; never override the prompt’s read because it feels uncomfortable.
- Sending AI-drafted QBR slides without your edit pass. Customers can tell when QBR content is generic; the work that builds trust is specific. Use AI for structure, you write the specifics.
- Pasting full customer PII or sensitive contract terms into a free-tier AI tool. Use paid plans with no-training and no-retention terms; verify your company’s AI policy before customer data flows through any tool.
What if you want to take this further?
Each prompt above takes inputs you paste in. The next move is connecting AI to the systems where CSM work happens.
Connectors are now standard
Claude, ChatGPT, and Grok all support connectors that let your AI read live data from your work tools (Gmail, Notion, GitHub, Asana, HubSpot, Stripe, and many more) instead of relying on you to paste context. For CSMs this means the AI can read your Gainsight or ChurnZero data, your Salesforce or HubSpot account history, your Gmail thread with the customer, or your Notion account brief library.
For Customer Success Managers, the connectors worth pairing with these prompts:
- Salesforce / HubSpot connector — reads account history, deal-stage data, and prior touch history for the QBR and renewal prompts.
- Gainsight / ChurnZero connector — if your CS platform exposes data via integration, AI reads health scores and activity directly.
- Gmail / Outlook connector — pulls customer correspondence for QBR and renewal prep.
- Notion connector — references your account brief templates and prior win stories.
- Calendly connector — for renewal and QBR scheduling context.
What are common questions about AI for customer success?
Should I use AI for QBR slides?
Use AI to structure. Do not let AI generate slides with content the customer will read. Customers detect generic QBR content within three slides. The QBR Deck Drafter prompt produces a structure; you write the customer-specific substance.
Will AI replace CSMs?
AI is changing what CSMs do. Health analysis, QBR prep, churn-signal detection: all compressing. Customer relationship management, executive sponsor cultivation, save conversation execution: still your work. CSMs who use AI for back-office work and spend their time on relationships become more valuable.
Which AI tool is best for CSMs?
Claude Pro for the judgment-heavy prompts (health reads, save plans, renewal prep) because of its discipline about not confirming optimism. ChatGPT for fast drafting work. Your CS platform’s built-in AI (Gainsight, Catalyst) for routine data work.
Is customer data safe in AI tools?
Paid Claude and ChatGPT plans do not train on inputs and do not retain content beyond the session. For enterprise CS work specifically, your security team should review the data flow before customer PII flows through any AI.
Should I tell customers I use AI?
Most CSMs do not announce specific back-office tools. Customers care about outcomes, not your tooling. If asked, be direct: AI helps me prepare; the read and the conversation are mine.
Can AI predict churn?
AI can surface patterns in the data you already have. AI cannot read whether your champion is being asked to find budget cuts or whether the new CTO has a vendor preference. Use AI for pattern surfacing; bring your relationship intelligence.
How long does it take to build the CSM-AI loop?
Three weeks. Start with the Account Health Read on your top 10 accounts. Add the Renewal Prep when your next renewal is 60 days out. Most CSMs settle into 4-5 of the seven prompts within a month.
The AI Prompt Library · $39
Customer success workflows, prompt-paved.
Soon to be 1000+ prompts in Notion organized by use case. The full CS section includes everything above plus prompts for advisory board recruitment, NPS follow-up segmentation, executive sponsor briefings, and CS-to-Product feedback memos. Plus prompts for every other field. Lifetime access.
Get Smarter About AI Every Morning
Free daily newsletter. Built for people who want to use AI well, not chase every model.
Free forever. Unsubscribe anytime.
Sources to read next?
- Anthropic prompt engineering documentation · official prompt design guide
- Gainsight on AI in Customer Success · industry research on CS practice
- Customer Success Association · professional development resources
- Anthropic: Introducing Connectors · context for the Salesforce, Gmail, Notion callout
- Forrester Customer Success Maturity research · evidence-based CS practice research
You might also like
- AI Prompt Library · the full library this post pulls from
- Best AI Prompts for Sales · the sales-side companion
- Best AI Prompts for PMs · for the PM partnership side
- How to Edit AI Out of Your Writing · the cleanup pass before customer-facing
- Prompt to Workflow: The AI Ladder · graduate prompts into saved workflows
- Best AI Prompts for Email Writing · for customer email volume
- Best AI Prompts for Meeting Notes · for customer call follow-up
Two ways to go further
The AI Prompt Library
1,000+ ready-to-use prompts for Claude, ChatGPT, and Gemini. Stop staring at a blank box.
Get it for $39 →2-Hour Live AI Crash Course
A private, beginner-friendly session across Claude, ChatGPT, Gemini, and the wider landscape.
Book for $125 →