Customer call transcripts → case studies + testimonial graphics + social proof posts

Sales calls and customer success calls carry the most valuable content most B2B companies record — and almost all of it dies in the call recording archive. A typical 30-minute customer call contains a complete case study, several quotable testimonials, and 3-5 data points your prospects need to see.

This is the workflow for systematically extracting that content into marketing assets, with consent and anonymization rules built in.

Get consent first — non-negotiable

Before you process any customer call recording for marketing content:

1. The customer must have opted into recording at the start of the call (Calendly recording disclosures or Zoom consent prompts satisfy this for the recording itself).
2. For external publication of quotes, you need explicit written consent. Email is fine. Standard language: "We would like to use the following quotes from our recent call in marketing materials. Are you comfortable with us publishing these attributed / anonymized [pick one]?"
3. If the customer says anonymize, you can still publish — but never name the customer or use details specific enough to identify them.
4. If the customer declines, the content stays internal. No exceptions.

Skipping this is the fastest way to burn customer trust and create legal exposure. Make it a checklist item before any extraction.

Step 1: Get a clean transcript with speaker labels

Same transcript hygiene as podcast repurposing. Speaker labels are non-negotiable for customer calls because the AI needs to know which lines are yours vs the customer's — only the customer's lines become quotes.

Step 2: Extract the story arc

Every good case study has 3 acts:

• Before: what was the situation that drove them to look for a solution? (problem framing)
• Change: what did they try, what changed, what was the inflection point? (transformation)
• After: what is different now? Quantify if possible. (result with numbers)

Scan the transcript for sentences that map to each act. The customer usually walks the rep through this naturally during discovery and during success calls. Your job is to extract and structure it.

Step 3: Pull verbatim quotes for graphics

Look for sentences that are:

• Complete — they stand alone without context
• Specific — name a number, a comparison, or a concrete outcome
• Conversational — read as the customer talking, not as marketing copy
• Quotable in under 20 words — testimonial graphics live or die on tight text

Examples of good extracted quotes:

• "We cut from 8 hours of content production a week down to under 1 hour."
• "I stopped paying for three tools and consolidated to one."
• "The Persona Brief made our AI content actually sound like us — I was skeptical until I saw it."

Bad extracted quotes (too marketing-y, too long, too vague):

• "Kompozy is the best AI content tool out there — game-changing for our team and we love every aspect of the platform."

Step 4: Decide attributed vs anonymized

Match what the customer agreed to in consent. Attributed (with name + company logo) is more powerful — every named case study is worth 3-5 anonymized ones in conversion impact. Anonymized still works if you keep the role specific ("SaaS founder, 50-person team") and the result-numbers concrete.

Step 5: Generate the marketing assets

From one consent-cleared customer call:

• 1 structured case study (1,000-1,500 words, before / change / after structure)
• 3-5 pull-quote testimonial graphics (one per quotable line)
• 1-2 LinkedIn posts framing the case study as a story
• 1 X thread walking through the customer journey
• 1 sales-enablement one-pager (PDF) for reps to share with similar-profile prospects

Anonymization checklist

When publishing anonymized case studies:

• Use role + industry, not name. "B2B SaaS founder" not "John Smith, CEO of Acme."
• Anonymize specific numbers if they would identify the customer ("from 100 customers to 1,000" if their actual numbers are unique enough to identify).
• Anonymize specific competitor mentions if the customer named one. "Switched from a competing scheduler" not "switched from Buffer."
• Strip any identifying jargon, industry codes, or geography unless generic.

Where to publish

• Named case studies → /case-studies on your site (high-trust, high-conversion surface)
• Pull-quote graphics → social (Instagram, LinkedIn, X — visual testimonials drive recall)
• LinkedIn story posts → LinkedIn (B2B audiences trust customer-story format)
• Sales one-pagers → internal CRM + outbound emails

Common failures

• Publishing without explicit consent. Legal exposure + customer trust burn.
• Editing quotes for clarity past the point of accuracy. The quote stops being the customer's.
• Generic anonymization that still identifies. "A SaaS founder in NYC running a 50-person team in fintech" is essentially named.
• Skipping the before/change/after structure. Reads as a testimonial pile instead of a story.
• Using AI-generated quotes presented as real customer quotes. This is fabrication. Hard rule: never.

Frequently asked questions

Related guides in AI Content Repurposing

• How to repurpose a podcast into 30+ pieces of social content — The 6-step methodology for fanning a 60-minute podcast episode into shorts, threads, carousels, a blog, and a newsletter.
• YouTube long-form to TikTok / Reels / Shorts: the complete repurposing workflow — Clip-detection, hook rewriting, platform-native captions, and scheduling cadence for YouTube long-form to short-form.
• Long-form to Instagram / LinkedIn carousels: the slide-extraction pattern — How to extract the right 6-8 ideas from a 2,000-word source and render them as brand-exact carousel slides.

Adjacent clusters

• AI Brand Voice & Persona — Without a Persona Brief, every AI output averages to the LLM default voice. This is the 5-section methodology that makes 100+ AI-generated posts feel like one human author wrote them.

← Back to AI Content Repurposing overview · Start a free trial → · See pricing