From Client Interview to Content: A Ghostwriter's Full Workflow

The End-to-End Process

Client interviews are raw material. Published content is the product. Here's the complete workflow from one to the other.

Phase 1: The Interview

Setup: Auto-recording enabled. Don't think about capture during the conversation.

Mindset: Your job isn't to take notes. It's to have a conversation that draws out authentic voice. Ask follow-ups. Let them tangent. The best quotes come from unexpected directions.

• •"Tell me more about that."
• •"What's an example?"
• •"How did that make you feel?"
• •"Why does that matter to you?"

Open-ended prompts. Let them fill the space.

Don't do: Interrupt to clarify for notes. You have the recording. Stay in the conversation.

Phase 2: Transcription

Automatic: Meeting ends. Transcription begins. No manual step.

What you get: Full verbatim transcript with speaker labels. Every word, every "um," every tangent.

Where it goes: Google Drive, client folder. Searchable. AI-accessible.

Time: Usually ready within minutes of the call ending.

Phase 3: Initial Review

• •Strong quotes (highlight or note timestamp)
• •Key themes
• •Surprising moments
• •Voice patterns worth capturing

Document your flags: Quick notes in a separate doc. "Great quote about competition at 23:00." "Interesting metaphor around 35:00."

This isn't full analysis. It's mapping the territory for later.

Phase 4: Voice Documentation

Build the voice document: A running reference of how this client talks.

• •Characteristic phrases ("Here's the thing," "At the end of the day")
• •Vocabulary preferences (technical terms, industry jargon)
• •Sentence structure patterns (short and punchy? long and flowing?)
• •Metaphor categories (sports, cooking, construction, war?)
• •Emotional expression style (understated? dramatic?)

Update after each interview. The voice document grows over time. By the third or fourth call, you have rich reference material.

Phase 5: AI-Assisted Analysis

Give Claude the transcript and voice document. Ask targeted questions:

• •"What are the three strongest quotes about [topic]?"
• •"How does this person describe [their product/customers/competition]?"
• •"What patterns do you see in how they express enthusiasm vs. concern?"
• •"Find examples of their characteristic phrases in use."

Claude mines the transcript for you. You get organized excerpts without re-reading everything.

Phase 6: Content Structuring

Outline first. Before writing: 1. Core message 2. Key supporting points 3. Evidence/quotes for each 4. Opening hook 5. Closing reinforcement

Map quotes to structure. Which interview moments support which points? Place them in your outline.

Check for voice consistency. Does the structure match how they think? Would they organize ideas this way?

Phase 7: First Draft

Write with voice document open. Reference constantly. Read their quotes aloud before writing.

Use their actual phrases. When possible, use exact language from transcripts. Not paraphrase. Their words.

Aim for authentic, not polished. First draft should sound like them talking. Edit for polish later.

AI assist option: "Write a section about [topic] in this person's voice. Here are example quotes: [quotes]. Maintain their style of [specific patterns]."

Review and revise. AI gets you 70%. Your editing gets to 95%.

Phase 8: Self-Edit

Read aloud. Does it sound like them? Would they say this?

Check against transcripts. Are you representing their views accurately? Did you capture their voice or impose yours?

Cut what's not them. Phrases you'd use. Structures that are your style. Be rigorous about authenticity.

Phase 9: Client Review

Send with context: "Here's the draft based on our conversation. I've tried to capture your voice—let me know if anything doesn't sound like you."

Invite specific feedback: "Are there phrases that feel off? Anything you'd never say?"

Iterate based on input. Their feedback trains your voice capture for next time.

The Transcript-Dependent Reality

This workflow depends on having full transcripts. Not summaries. The actual words.

• •Quote mining (no quotes to mine)
• •Voice documentation (no examples to document)
• •AI analysis (nothing substantial to analyze)
• •Client verification (no source to check against)

Summaries give you information. Transcripts give you raw material.

Every step of this workflow relies on access to what was actually said.

Tools for the Workflow

Recording: Calendar-triggered auto-start. Never miss capturing a conversation.

Transcription: Deepgram or similar. Accurate, affordable, fast.

Storage: Google Drive client folders. Searchable, AI-accessible.

AI assist: Claude with Drive integration. Reads transcripts directly.

Writing: Whatever you prefer. The magic is in the preparation, not the editor.

The Result

Content that sounds like the client wrote it themselves.

Not because you're a mind reader. Because you have a system that captures their voice and keeps it accessible throughout the writing process.

The workflow takes discipline to establish. But once running, it produces authenticity that's almost impossible to achieve from memory and notes alone.

Start with capture. Build from transcripts. Let the client's actual words guide every step.