Claude-optimized prompt structure shaped for chatgpt.
# Task
Craft a response to a buyer objection. Use ONLY the provided facts and proof points.
# Inputs- Objection: {objection}- Buyer context: {buyer_context}- Product: {product}- Proof points: {proof_points}- Tone: {tone}- Constraints: {constraints}# Anti-hallucination rules1. Do not invent case studies, stats, or customer names not in the proof points.
2. Do not promise capabilities not described in the product context.
3. If you lack evidence to fully counter the objection, say so and suggest what evidence would help.
# Response framework (use this structure)1. **Acknowledge** (1 sentence): Validate the concern. Show you heard them. Never dismiss or minimize.
2. **Reframe** (1-2 sentences): Shift the framing — from cost to ROI, from risk to cost-of-inaction, from complexity to time-to-value.
3. **Evidence** (1-2 sentences): Specific proof point. Name, number, or concrete result. Not "many customers say..."
4. **Bridge** (1 sentence): Connect back to what matters to THIS buyer. Tie to their stated pain or goal.
# Stall detection
If the objection is a stall ("I'll think about it", "Send me info", "Check back next quarter"), flag it as [LIKELY STALL] and provide:
- A respectful unstalling question that surfaces the real concern.
- A soft commitment ask (specific date, intro to another stakeholder, or a smaller next step).
# Confidence rating
Rate your response:
- **Strong**: Multiple proof points directly address this objection.
- **Moderate**: Partial evidence — response is credible but could be stronger.
- **Weak**: Limited evidence — flag what's missing.
# Output1. Response using the framework above
2. Short talk-track version (3-4 sentences, conversational)
3. Stall handling (if applicable)
4. Follow-up question to deepen the conversation
5. Confidence rating with explanation
claudegrounded_factsmulti_perspective
claude variant
Claude-optimized prompt structure shaped for claude.
<context><objection>{objection}</objection><buyer_context>{buyer_context}</buyer_context><product>{product}</product><proof_points>{proof_points}</proof_points><tone>{tone}</tone><constraints>{constraints}</constraints></context><task>Craft a response using ONLY the provided facts and proof points.</task><instructions>
No invented stats or customer names. If evidence is insufficient, say so.
Framework: Acknowledge (validate) > Reframe (shift framing) > Evidence (specific proof) > Bridge (connect to buyer's pain).
If stall detected ("I'll think about it"): flag [LIKELY STALL], provide unstalling question and soft commitment ask.
Rate confidence: Strong / Moderate / Weak with explanation.
</instructions><output>
Return: structured response, short talk-track (3-4 sentences), stall handling (if applicable), follow-up question, confidence rating.
</output>
geminigrounded_factsmulti_perspective
gemini variant
Claude-optimized prompt structure shaped for gemini.
Objection: {objection}
Buyer context: {buyer_context}
Product: {product}
Proof points: {proof_points}
Tone: {tone}Constraints:{constraints}
Rules: only provided facts, no invented stats. If evidence insufficient, say so.
Framework: Acknowledge > Reframe > Evidence > Bridge (to buyer's pain).
If stall detected: flag [LIKELY STALL], provide unstalling question and soft commitment ask.
Rate confidence: Strong / Moderate / Weak.
Based on the entire content above, return: structured response, talk-track (3-4 sentences), stall handling (if applicable), follow-up question, confidence rating.
Related templates
More templates in the same working category.
Email Reply
Draft a professional email reply that matches the thread tone, addresses every point raised, and drives toward a clear next step.