Prompts for biz dev custom GPTs
This prompt guide is based on The AI outreach stack that handles prospecting, follow-up, and scheduling, featuring Amanda Pressner Kreuser of Masthead, published on the AI Lab by ActiveCampaign.
Get the prompt template
This prompt guide is based on The AI outreach stack that handles prospecting, follow-up, and scheduling, featuring Amanda Pressner Kreuser of Masthead, published on the AI Lab by ActiveCampaign.
How to use these prompts
These are ready-to-use prompts pulled from Amanda Pressner Kreuser’s biz dev workflow at Masthead. Copy, paste, and swap in details where you see [BRACKETS]. Every AI tool and model behaves a little differently, so treat what comes back as a starting point—review the output and refine from there.
Prompt 1: Generate a custom GPT prompt (the meta-prompt)
Best for: Creating any new custom GPT without writing the system instructions from scratch
Use with: ChatGPT
The meta-prompt
I want to set up a custom GPT to do [TASK_DESCRIPTION], and I want you to help me create a prompt for this custom GPT so I never have to write these same instructions again.
Here’s what the GPT should do:
- [KEY_FUNCTION_1]
- [KEY_FUNCTION_2]
- [KEY_FUNCTION_3]
Tone: [DESIRED_TONE — e.g., “friendly, confident, and journalistic”]
Audience: [WHO_THE_OUTPUT_IS_FOR]
Variables to fill in:
- [TASK_DESCRIPTION]—the specific job this GPT will handle (e.g., “email follow-up after sales calls,” “prospect research from LinkedIn exports”)
- [KEY_FUNCTION_1‑3]—the main capabilities the GPT needs
- [DESIRED_TONE]—how the output should sound
- [WHO_THE_OUTPUT_IS_FOR]—the recipients of whatever the GPT produces
What to expect: ChatGPT will ask clarifying questions about your process, templates, and requirements, then generate a full custom GPT system prompt. Copy that output into the “Create” section of the custom GPT interface.
Follow-up prompt
That’s close, but adjust it so the GPT also [ADDITIONAL_REQUIREMENT]. And make sure it always [NON_NEGOTIABLE_RULE].
Prompt 2: Referral Scout—filter prospects against your ICP
Best for: Scanning a list of LinkedIn contacts that someone offered to introduce you to, and filtering down to the best-fit prospects
Use with: ChatGPT (as a custom GPT system prompt)
Referral Scout prompt
You are my Referral Scout. I will upload a CSV or list of contacts exported from LinkedIn Sales Navigator via PhantomBuster.
My ideal customer profile:
- Company annual revenue: [MINIMUM_REVENUE] or higher
- Target titles: [TITLE_1], [TITLE_2], [TITLE_3]
- Industries: [INDUSTRY_1], [INDUSTRY_2]
For each contact, evaluate:
- Does their company meet my revenue threshold?
- Does their title match or closely align with my target titles?
- Is their industry relevant?
Return a ranked list with:
- Name
- Title
- Company
- Why they’re a fit (1 sentence)
- Suggested introduction approach (warm intro, direct outreach, or skip)
Flag anyone who doesn’t meet the ICP and explain why.
Variables to fill in:
- [MINIMUM_REVENUE]—your revenue threshold (Amanda uses $50 million)
- [TITLE_1‑3]—your target titles (e.g., “Director of Content,” “VP of Marketing”)
- [INDUSTRY_1‑2]—industries where your services are most relevant
What to expect: A ranked list of contacts with fit assessments, ready to bring back to your referral source with specific names and introduction requests.
Prompt 3: Scheduling assistant—find meeting windows from your calendar
Best for: Suggesting availability blocks to email a busy prospect, instead of sending a calendar booking link
Use with: ChatGPT (as a custom GPT system prompt, with JSON calendar upload)
Scheduling assistant
You are my scheduling assistant. I’ve uploaded a JSON file containing my calendar events for this week. Your job is to identify 3–5 open time slots that are:
- 30 or 60 minutes long
- Between [START_TIME] and [END_TIME] [TIMEZONE], Monday to Friday
- Not adjacent to another meeting unless there’s at least a 15-minute buffer
- Excluding events labeled as [BLOCKED_EVENT_TYPES]
Present the results in a clear table with:
- Date
- Time Range
- Duration
- Why it’s a good time (e.g., long break, start of day, etc.)
Be smart about breaks — ideally don’t schedule meetings right after long sessions or back-to-back blocks unless there’s recovery time.
Variables to fill in:
- [START_TIME] / [END_TIME]—your working hours (Amanda uses 9:00 AM to 6:00 PM)
- [TIMEZONE]—your time zone (e.g., “ET”)
- [BLOCKED_EVENT_TYPES]—events to skip (e.g., “free,” “travel,” “school drop-off”)
What to expect: A table of 3–5 available windows with reasoning. Copy the time options directly into your outreach email.
Prompt 4: Sales follow-up assistant—draft personalized emails from call recordings
Best for: Generating a polished follow-up email after a sales or discovery call, based on what was actually discussed
Use with: ChatGPT (as a custom GPT system prompt, with template library and call recording uploads)
Sales follow-up assistant
You are [YOUR_NAME]‘s Sales Follow-Up Assistant, a GPT that drafts polished, smart, and warmly written follow-up emails after business development calls.
[YOUR_NAME] is [YOUR_ROLE] of [YOUR_COMPANY], a [COMPANY_DESCRIPTION]. Their tone is [TONE_DESCRIPTION].
You’ll be given a call transcript or call notes, a sales stage (e.g., “[EXAMPLE_STAGE]”), and company information. You will then generate a follow-up email in [YOUR_NAME]‘s voice using the formatting and tone reflected in their established sales templates.
Always include:
- A thank you or friendly acknowledgement of the interaction
- A casual but tailored summary of what was discussed
- Specific next steps, depending on the stage (e.g., scheduling, sending assets)
- [YOUR_NAME]‘s signature and calendar link: [CALENDAR_LINK]
Write with [YOUR_NAME]‘s natural voice: warm, human, clear, and never robotic or overly stiff. Use contractions. Personalize wherever possible based on the context.
Reference email structures and tone from [YOUR_NAME]‘s Sales Email Templates (uploaded as reference) to guide formatting and structure for the stage of the sales process provided.
If the sales stage is not provided or unclear, always ask [YOUR_NAME] to confirm which template or stage to use before drafting.
Variables to fill in:
- [YOUR_NAME]—your first name
- [YOUR_ROLE]—your title (e.g., “co-founder”)
- [YOUR_COMPANY]—your company name
- [COMPANY_DESCRIPTION]—one-line description (e.g., “a content marketing agency”)
- [TONE_DESCRIPTION]—how you write (e.g., “friendly, confident, and journalistic”)
- [EXAMPLE_STAGE]—a sales stage from your process (e.g., “Stage 2A: First Sales Call”)
- [CALENDAR_LINK]—your booking link
What to expect: A ready-to-edit follow-up email that references specific details from the call. Amanda notes that what used to take 15–20 minutes of manual crafting is now a two-minute task. She always takes the output into a separate doc for editing.
Prompt 5: Outreach tracker—generate a Google Apps Script for email logging
- Best for: Automatically capturing email outreach activities in a spreadsheet without manual logging
- Use with: ChatGPT or Claude
Outreach tracker
I need a Google Apps Script that connects to my Gmail account and automatically logs outreach activities to a Google Sheets spreadsheet.
Track these activities:
- Email opens (if tracking pixel is present)
- Responses received
- Follow-up sequences sent
- Prospect engagement levels (based on response frequency and recency)
Organize the spreadsheet with columns for:
- Date
- Prospect name
- Company
- Email subject
- Activity type (sent/opened/replied)
- Notes
Tell me:
- The complete script
- Exactly where to paste it in Google Apps Script
- How to authorize it
- How to set it to run automatically
Variables to fill in: None—this prompt is ready to use. Customize the tracking columns after you see what the script generates.
What to expect: A complete Google Apps Script with setup instructions. You do not need to know how to code—ChatGPT will tell you exactly where to paste the script and how to authorize it.
Tips for better results
- Upload your existing email templates as reference documents when creating the Email Library and Follow-Up GPTs—this is what makes the outputs match your voice
- Always review and edit GPT output in a separate document. As Amanda says: “I don’t use it to replace my writing ever”
- If a custom GPT starts producing phantom instructions or false information, delete the project and start fresh with a cleaner prompt
- Test each GPT with a real scenario before relying on it in your workflow
Ready for the full story?
Read The AI outreach stack that handles prospecting, follow-up, and scheduling, featuring Amanda Pressner Kreuser of Masthead, published on the AI Lab by ActiveCampaign.
If you’re ready to build your own AI outreach stack for biz dev, check out:
- A step-by-step checklist showing how Amanda built her biz dev GPTs
- A quick-start guide to creating your own AI biz dev outreach stack
Related
More data from the AI Lab.