A guide to creating a podcast-booking custom GPT

The build (minimum viable version)

Step 1: Create the custom GPT and feed it your foundation—30 minutes

Go to ChatGPT’s custom GPT builder and start a new project. Paste in your services page, bio, and 2–3 past project descriptions. Define your ideal client profile and list which industries or topics are off-limits. Add a few personal details that make you interesting as a guest—hobbies, values, personal story.

Quick version: Don’t wordsmith your inputs. Copy-paste raw text from your website. You can refine the language later. Use ChatGPT’s voice-to-text feature to talk through personal details instead of typing them.

You’re doing it right if: You can ask the GPT ​“What kind of podcasts would be a good fit for me?” and it gives a specific, accurate answer that reflects your actual services and positioning.

Step 2: Set up the vetting and pitching workflow—30 minutes

Tell the GPT its job is to evaluate podcast opportunities and write pitches. Configure it to follow this sequence: (1) you paste a podcast description, (2) it decides whether the show is a fit and explains why, (3) if it’s a yes, it asks for 3–5 recent episode titles and descriptions, (4) it asks whether outreach is via email or form, and (5) it generates the appropriate pitch.

Quick version: Start with email pitches only. Skip the form-response branch for now and add it after your first few successful pitches.

You’re doing it right if: You paste a real podcast description, the GPT correctly identifies whether it’s a fit, and the pitch it generates references specific topics that match the show’s audience.

Step 3: Test with a real podcast and refine—30 minutes

Find a podcast you’d genuinely like to pitch and run it through the workflow end to end. Evaluate the pitch quality: Is the tone right? Does it accurately represent what you’d talk about? Give the GPT direct feedback on anything you’d change—wording, tone, topic framing.

Quick version: Test with one podcast, not five. Focus on getting one good pitch before scaling. If the pitch is 80% there, note what to fix and move on.

You’re doing it right if: The GPT produces a pitch you’d feel comfortable sending with only minor edits, and it improves noticeably after you give it one round of feedback.

You’re done when…

You can paste a podcast description into your custom GPT and get back: 

1. A clear yes/​no recommendation with reasoning, and 
2. A personalized pitch that represents your brand and speaks to the show’s specific audience. The pitch should require minimal editing before sending.

Level up later

When you’re ready to go deeper:

• Add form-response handling so the GPT can fill out podcast guest application forms
• Set up a pitch summary step for tracking all outgoing pitches in a shared doc
• Hand the GPT off to a team member and establish a weekly feedback loop
• Update the GPT regularly with new client stories and lessons from past appearances
• Use ChatGPT for bulk podcast vetting—paste several podcast descriptions at once and let it filter for fit