Skip to content
You commented “FEEDBACK”: here it is

Customer Feedback to Action Plan: Three-Prompt System

Run three prompts on raw customer feedback from any source and get a prioritized product and messaging improvement plan ready for your next sprint.

Free Claude skill · copy-paste ready · built by the BrainVaultAI team

Feedback is only useful if you extract the signal from the noise. Raw NPS comments, support tickets, and Intercom conversations are unstructured. Running them through a consistent analysis system turns them into decisions. This is a three-prompt system. Prompt 1 ingests and classifies raw feedback. Prompt 2 identifies the top themes and their root causes. Prompt 3 translates the findings into a prioritized action list, split between product, support, and messaging changes. ### How to use

Run Prompt 1 with your raw feedback batch. Paste Claude's output directly into Prompt 2. Then paste Prompt 2's output into Prompt 3. The whole cycle takes about ten minutes for a batch of 20-50 feedback items. Run monthly or after any major release. The output from Prompt 3 goes straight into your product backlog or your next team meeting.

Prompt 1: Feedback Ingestion & Classification
Classify the following customer feedback items. For each item:
- Sentiment: Positive / Negative / Mixed / Neutral
- Category: Bug / Feature Request / UX Friction / Pricing / Support / Praise / Churn Signal
- Signal strength: High (clear specific pain) / Medium / Low (vague)
- One-line summary of the core point Feedback items:
[PASTE 10-50 FEEDBACK ITEMS, can be NPS verbatims, support tickets, reviews, emails] Return as a table. Flag any item that appears to be a churn signal in red text.
Prompt 2: Theme & Root Cause Extraction
Using the classified feedback from the previous step, identify: 1. Top 3-5 themes (grouped patterns across multiple items, not just the most common word)
2. For each theme: - What customers are explicitly saying - What they probably mean (the underlying need or frustration) - How many items map to it (approximate) - Whether it's a product problem, a messaging problem, or a support problem
3. The most alarming single data point and why
4. The most encouraging single data point and why Do not just count words. Find the patterns that matter.
Prompt 3: Action Prioritization
Based on the themes above, produce a prioritized action list. My constraints:
- Team size: [SIZE]
- Current sprint capacity: [STORY POINTS or HOURS]
- Revenue stage: [ARR or MRR range] Return three lists: PRODUCT ACTIONS (ranked by impact/effort ratio):
- [Action] | Impact: H/M/L | Effort: H/M/L | Rationale MESSAGING FIXES (things we should change in how we explain the product):
- [What to change] | [Where, onboarding / website / emails / docs] | [What to say instead] SUPPORT / PROCESS FIXES (ops changes to reduce repeat issues):
- [Action] | Owner: [who should own this] | Timeline: [this week / this month / backlog] Top recommendation: the single highest-leverage thing to act on this week.

Want the editable version?

We'll send the editable version of this asset, plus a short series of our highest-leverage AI plays for operators. No spam. Unsubscribe in one click.

Free. No spam. Unsubscribe in one click. By submitting you agree to receive emails from BrainVaultAI.

Go further

Want the full founder vault? 51 prompts, ready to run.

51 production-grade prompts (Operator, Climber, Independent)
Copy-paste ready, no theory
Lifetime updates
Today$47

One-time $47. Instant access. No subscription.

More free AI guides in the full library · follow @brainVaultAI