The guide on building a five-GPT BizDev system
This guide is based on the How AI Maturity Levels Transform Your Marketing Work Week webinar, published on the AI Lab by ActiveCampaign.
Get the quick-start guide
What will I accomplish with this guide? By the end of this guide, you’ll have a five-CustomGPT system that handles prospecting, outreach, and follow-up for a solo BizDev pipeline. It’s based on the workflow Amanda Pressner Kreuser walked through in this webinar, the same setup that took her from spending 20+ hours a week on manual sales tasks to recovering at least four to five hours every week.
Before you start, you’ll need:
- A ChatGPT Plus account (CustomGPTs require a paid plan)
- A LinkedIn Sales Navigator subscription with a saved Ideal Lead Profile search
- A third-party export tool that respects Sales Navigator’s terms of service (Amanda uses Phantom)
- An AI meeting note-taker (Otter, Fireflies, or similar) that produces transcripts you can upload
- A library of 10–15 of your best follow-up and outreach emails saved as a single document
- A Google account if you want to use the activity-tracker spreadsheet phase
Quick reference
- Total time: 4–6 hours to set up all five GPTs; ongoing use is minutes per task
- Tools needed: ChatGPT (CustomGPTs), LinkedIn Sales Navigator, Phantom (or equivalent), AI note-taker, Google Sheets + Apps Script
- Key output: A repeatable BizDev pipeline where prospect research, outreach drafting, scheduling, follow-ups, and activity tracking each run through a dedicated GPT
Watch this section
For full context on the following topics, watch these sections of the webinar:
- The five-GPT overview and why each one exists — [08:03]–[12:05]
- Live referral-scout demo in Sales Navigator + Phantom + ChatGPT — [13:02]–[17:30]
- Live follow-up GPT demo using a real sales call — [17:30]–[19:05]
The workflow
Phase 1: Map your manual sales tasks before you build anything
After this phase, you’ll have: a list of the repetitive BizDev tasks that are good candidates for a CustomGPT, plus the source documents each one needs.
- Audit your last two weeks: write down every BizDev task you did manually, how long it took, and how often it repeats.
- Group tasks by trigger: prospecting, outreach drafting, scheduling, post-call follow-up, monthly tracking. Those are Amanda’s five buckets; use whatever fits your pipeline.
- Pull your best emails into one doc: Amanda kept 10–15 of her go-to follow-up and intro templates in a Word doc. That doc becomes the training input for the writing GPTs.
- Document your ideal customer profile (ICP): company size, industry, seniority, decision-maker title, all written out in plain English so a GPT can reason against it.
I had to create every single sales process from scratch, not knowing what I was doing. Before workflows, I was spending at least 20 hours a week, if not more, on stuff like prospect research, outreach, and follow ups.
Phase 2: Build the referral scout
After this phase, you’ll have: a CustomGPT that takes a list of someone’s LinkedIn connections and returns the top 10 people you should ask for an intro to.
- Save an Ideal Lead Profile search in Sales Navigator: filter by company size (Amanda uses 500+), function (marketing), and seniority. Save it so you can reuse it for any contact’s connection list.
- Set the search to second- and third-degree connections of one specific contact: under All Filters, narrow to connections of, for example, the person you’re hoping to get a referral from.
- Export the result with a Sales Navigator-compliant tool: Amanda uses Phantom sparingly, capping at 250 leads per pull and respecting the 2,500/day platform ceiling.
- Create a CustomGPT called “Referral scout”: train it on your ICP description from Phase 1.
- Tell the GPT what to return for each match: name, company, one-line justification, and the LinkedIn URL.
- Use it: upload the exported CSV, ask for the top 10 referral picks, and send the shortlist to your contact with a pre-written intro request.
Phase 3: Build the outreach + follow-up writer
After this phase, you’ll have: a CustomGPT that drafts a personalized follow-up after every sales call by combining the call transcript with your existing email library.
- Create a CustomGPT called “Smart warm sales follow-up”: load your 10–15 best emails into the GPT’s training/knowledge files so it has a voice to mirror.
- Add your standard links to the training: calendar URL, case studies, capabilities deck, anything you reference often, so the GPT inserts them automatically instead of asking you each time.
- Define the input format: the GPT expects an AI-meeting transcript (Otter, Fireflies, or equivalent) plus optional notes on what stage of outreach this is.
- Define the output format: a draft that pulls specific points from the call and matches one of your saved templates by stage.
- Use it: upload the post-call transcript, copy the draft into your email client, and spend ~60 seconds making it sound like you before sending.
Phase 4: Add the scheduling assistant
After this phase, you’ll have: a way to drop three to five real availability windows into a follow-up email instead of sending a calendar link.
- Connect ChatGPT to your calendar: use the ChatGPT calendar connector (or equivalent) so it can read your availability.
- Define the request: tell the GPT to suggest three to five 30-minute windows over the next two weeks during your working hours.
- Drop the windows into the follow-up draft: Amanda prefers this over a Calendly link for senior contacts because it shows she’s already done the scheduling work.
Phase 5: Build the activity tracker
After this phase, you’ll have: a Google Sheet that auto-logs your outreach activity each month and a GPT that surfaces your best prospects.
- Describe what you want to ChatGPT in plain English: Amanda asked for a Google Apps Script that logs her outreach into a spreadsheet. She doesn’t code, so the GPT wrote the script and told her where to paste it.
- Paste the generated code into Google Apps Script: open your tracker sheet, go to Extensions → Apps Script, and follow ChatGPT’s placement instructions.
- Run a test month: confirm the script writes the right rows; refine the prompt if it doesn’t.
- Build a “best prospects” GPT on top of the sheet: at month-end, upload the sheet to a GPT trained to flag your warmest leads based on activity recency and engagement.
- Use it for your monthly partner review: Amanda and her co-founder use the output as the agenda for their one-on-one outreach review.
I just described what I needed into the regular ChatGPT and then ChatGPT wrote the code and told me exactly where to paste it within the Google environment. So that’s the power of treating AI like an on-demand tutor.
Related
More data from the AI Lab.