Skip to content This library is free, offered as dana. If it helps your work, donate here →
penguin tree ai
← back to libraryDonate here

This is one of 3,622 free prompts in the penguin tree ai library. Sign up once and the whole library opens, with new prompts by email.

open the library →

Customer testimonial generator

Transforms raw customer interview notes, call transcripts, or survey responses into a set of polished, on-brand testimonials formatted for 6 distinct channels — website quote, case study pull-quote, sales deck slide…

download .md

Purpose

Turn messy, unstructured customer feedback — interview notes, call transcripts, survey verbatims — into a set of clean, quotable testimonials tuned for specific channels, each anchored to a concrete outcome and written in the customer's own register. Solves the recurring bottleneck where great customer stories die in a Notion doc because nobody has time to shape them into publishable copy.

When to Use

Invoke this skill when the user asks to:

  • Turn customer interview notes or a call transcript into publishable testimonials
  • Pull quotable lines out of raw survey responses or NPS verbatims
  • Create testimonial variations for a website, sales deck, or social post
  • Polish a rough customer quote without making it sound fake or corporate
  • Generate a testimonial in multiple lengths for different placements
  • Extract the strongest proof points from a customer success story

Inputs Required

Before starting, collect:

  • Raw feedback: The interview notes, transcript excerpt, or survey answers (paste as-is)
  • Customer identity: Name, job title, and company (or note if it must stay anonymous/attributed by role only)
  • Product/service: What the customer bought or used
  • Primary outcome: The main result the customer achieved (metric if available)
  • Approval constraints: Any claims that can't be published, or metrics that must be softened (e.g., "3x" → "significant increase")
  • Target channels: Which of the 6 outputs are needed (default: all)
  • Voice preference: Should quotes preserve casual/spoken tone, or read more polished?

If customer identity or voice preference is missing, default to role-and-industry attribution and a lightly-polished-but-natural voice, and flag the assumption.

Instructions

Step 1: Mine the Raw Feedback for Signal

Read the raw input and extract, without inventing anything:

  • The problem before: What pain or friction existed pre-purchase
  • The turning point: Why they chose this product / what changed
  • The specific outcome: Quantified if possible; qualitative if not
  • Emotional language: The customer's own vivid phrases, adjectives, and metaphors — these are gold and should be preserved verbatim where possible
  • Objections overcome: Any hesitation they mention that others share

Rule: every testimonial must trace to something the customer actually said or implied. If a quote requires a claim not supported by the input, flag it as [NEEDS CUSTOMER CONFIRMATION] rather than fabricating it.

Step 2: Identify the Strongest Proof Point

Rank the extracted outcomes by persuasive weight:

  1. Specific quantified result ("cut onboarding from 3 weeks to 4 days")
  2. Named business impact ("we hit our Q3 pipeline target for the first time")
  3. Emotional/relief statement ("I stopped dreading Monday reporting")
  4. Comparison to prior solution ("nothing else we tried came close")

Lead every channel variation with the highest-ranked proof point available.

Step 3: Generate 6 Channel Variations

Produce one testimonial per requested channel. Each preserves the customer's voice but is shaped for its placement:

1 — Website Hero Quote (15–25 words) Punchy, single-idea, outcome-forward. The line that could sit next to a headshot on a landing page. No hedging, no run-ons.

2 — Case Study Pull-Quote (30–50 words) Slightly longer, includes the before/after contrast. Reads as a natural break inside a longer story.

3 — Sales Deck Slide Quote (10–18 words) Tightest possible version. Must land in one breath when read aloud on a call. Metric-forward if a metric exists.

4 — Social Proof Snippet (20–40 words) Conversational, shareable, feels human — as if the customer posted it themselves. Can retain a spoken cadence.

5 — Review-Site Paste (60–90 words) First-person, structured as problem → solution → result. Reads like an authentic G2/Capterra review with a light star-rating cadence.

6 — Long-Form Narrative Testimonial (120–180 words) Full arc in the customer's voice: the situation, the decision, the outcome, and a forward-looking closing line. Suitable for a quote block, email, or PR use.

For every variation, append the attribution line: — [Name], [Title], [Company] (or role-based attribution if anonymous).

Step 4: Flag Compliance and Authenticity Risks

Below the variations, list:

  • Any claim marked [NEEDS CUSTOMER CONFIRMATION]
  • Any metric that was softened per approval constraints
  • Any line that leans stronger than the raw input supports, with a note

Step 5: Recommend the One to Lead With

Name the single strongest variation and where to deploy it first, with one sentence of reasoning.

Output Format

  • Structure: 6 clearly labeled channel variations, each with word count and attribution, followed by a Compliance Flags section and a Lead Recommendation
  • Length: 500–800 words total
  • Format: Labeled blocks per channel (not prose); each quote set off clearly
  • Sections: Extracted Proof Points (brief) → 6 Channel Variations → Compliance Flags → Lead Recommendation
  • Tone: Authentic to the customer's voice — human, specific, never corporate-testimonial cliché ("game-changer," "took our business to the next level" are banned unless the customer actually said them)

Quality Checklist

Before delivering output, verify:

  • [ ] All requested channel variations are present with correct word-count ranges
  • [ ] Every quote traces to something in the raw input — nothing fabricated
  • [ ] Any unsupported or softened claim is flagged, not silently inserted
  • [ ] The customer's own vivid phrases are preserved where they exist
  • [ ] Attribution line is present on every variation
  • [ ] A single lead recommendation is named (no hedging)

Notes

  • The most common failure in testimonial writing is over-polishing until it sounds like the marketing team wrote it. Preserve one or two "imperfect" human phrases per long-form version — they read as more credible than clean copy.
  • If the raw input is thin, extract fewer strong testimonials rather than padding weak ones. A single sharp 18-word quote beats six generic ones.
  • Metrics are the highest-converting element in a testimonial. If the input has a number, surface it in at least three of the six variations.
  • Always advise the user to route final testimonials back to the customer for sign-off before publishing — attribution changes the legal and reputational stakes.

How to use this skill

This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.

Installing in Claude Code:

  • ~/.claude/skills/<folder-name>/SKILL.md — auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)
  • <your-project>/.claude/skills/<folder-name>/SKILL.md — auto-loads only when working in that project (recommended for project-specific skills)

Restart Claude Code or open a new session to register.

Using in Claude.ai, ChatGPT, Gemini, or other AI tools:

  1. Drag this file into a new chat, or copy and paste its contents
  2. Send your task or request as your next message — the AI will apply the skill where it fits

For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills