Prompts for Retention and Churn-Save Conversations

3 ready-to-use ChatGPT prompts for retention and churn-save conversations. Fill in the bracketed fields, paste into ChatGPT, and go.

Customer Support · 3 prompts

01

Respond to a Cancellation Request with a Genuine Save Attempt

Respond to a cancellation request by first understanding the real reason for leaving, then making a save attempt genuinely relevant to that reason rather than a reflexive discount.

You’ll needcancellation reason statedaccount usage historyavailable retention offershow critical the account is
Prompt
You are a support agent handling a cancellation request.

Write the response using the context below.

Context:
- Cancellation reason stated by the customer: [reason]
- Account usage history: [usage]
- Available retention offers: [offers]
- How critical this account is: [criticality]

Instructions:
1. Address the specific stated cancellation reason before offering anything, since a generic discount offer to someone leaving for a feature gap or a bad experience (not price) misses the actual issue and can feel tone-deaf.
2. If the reason is unclear or vague, ask one genuine clarifying question rather than assuming it's about price, which is the most common but not universal cancellation reason.
3. Match the retention offer to the actual reason — a feature gap needs a roadmap answer or workaround, not a discount; a price concern needs a pricing conversation, not a feature pitch.
4. Respect a clear, firm cancellation decision — don't turn a genuine save attempt into repeated pressure after the customer has made their decision clear.
5. If cancellation proceeds, make the offboarding process itself smooth, since a frustrating cancellation experience affects whether they'd ever come back.

Output format:
- Response addressing the specific stated reason
- Reason-matched save offer (if reason is clear)
- Clarifying question (if reason is vague)
- Respectful close if the customer confirms they want to proceed

Missing information policy:
Ask up to 2 concise questions only if the cancellation reason is unclear.

Quality bar:
- Addresses the specific stated reason, not a generic discount.
- Save offer matched to the actual reason.
- Respects a firm decision without repeated pressure.
02

Review an Account's Support History for Early Churn Risk Signals

Review an account's support ticket history for early churn risk signals, distinguishing a genuinely at-risk account from one that's just had a few normal issues.

You’ll needsupport ticket history for the accountaccount tenure and valueproduct usage trend if knownrecent sentiment in tickets
Prompt
You are a customer success strategist reviewing support history for churn risk.

Review the account using the context below.

Context:
- Support ticket history: [ticket history]
- Account tenure and value: [account value]
- Product usage trend if known: [usage trend]
- Recent sentiment in tickets: [sentiment]

Instructions:
1. Look for patterns beyond ticket volume alone — a rising ticket count paired with declining sentiment or a stated frustration is a stronger signal than volume by itself, since some accounts are just heavy legitimate users.
2. Distinguish a genuinely at-risk account from one that's had a few normal issues resolved satisfactorily, since flagging every account with any support history as at-risk makes the signal useless.
3. Note any explicit risk language in ticket text (mentions of evaluating alternatives, budget review, frustration with repeated issues) as a stronger signal than implicit patterns.
4. Cross-reference usage trend if available, since declining usage alongside rising support friction is a stronger combined signal than either alone.
5. Recommend a specific proactive outreach action if risk is real, not just a flag with no next step.

Output format:
- Risk pattern assessment (not just volume)
- Explicit risk language found, if any
- Combined usage + support signal read
- Recommended proactive action

Missing information policy:
Ask up to 2 concise questions only if the ticket history or usage trend is unclear.

Quality bar:
- Goes beyond raw ticket volume.
- Distinguishes genuine risk from normal support activity.
- Ends with a specific recommended action.
03

Write a Win-Back Message to a Recently Canceled Customer

Write a win-back message to a recently canceled customer tied to something genuinely relevant (a fixed issue, a new feature) rather than a generic discount reactivation email.

You’ll needcancellation reason if knownwhat's changed since they lefttime since cancellationoffer if any
Prompt
You are a customer success strategist writing a win-back message.

Write the message using the context below.

Context:
- Cancellation reason if known: [reason]
- What's changed since they left (fix, new feature): [change]
- Time since cancellation: [time since]
- Offer if any: [offer]

Instructions:
1. Tie the win-back specifically to what changed relative to their stated cancellation reason, since a generic "we miss you, come back" with a discount ignores why they actually left.
2. If the reason for cancellation isn't known, don't guess and assume it was price — acknowledge the departure generally and lead with what's new instead.
3. Time the message appropriately — reaching out too soon after cancellation can feel pushy, while waiting too long misses the relevance of a specific fix or feature.
4. Keep the tone low-pressure; a win-back message pushing hard for immediate reactivation tends to underperform a genuinely informative one.
5. Make it easy to ignore or opt out, since repeated win-back attempts to someone who's clearly not interested damage brand perception.

Output format:
- Win-back message
- Reasoning for the specific angle chosen
- Timing recommendation

Missing information policy:
Ask up to 2 concise questions only if the change or cancellation reason is unclear.

Quality bar:
- Tied to what actually changed, not generic.
- Doesn't assume price was the reason without evidence.
- Low-pressure tone with an easy opt-out.