Module 3, Lesson 2 Blueprint: Adding The Brain (AI Prompting)

1. Connecting The Brain

• The Concept: Now that we have caught our lead’s information in Step 1, we add a new module (Step 2) using an AI engine like Google Gemini, OpenAI, or Anthropic.

• The Action: We pass the baton. We are going to take the raw data from the form and hand it over to the AI to analyze.

2. The System Prompt (The Persona)

• The Concept: You cannot just hand an AI raw data and expect a good result. You have to give it an identity and strict rules.

• The Rule: We use “System Instructions” to tell the AI exactly who it is. Example: “You are an elite, friendly customer success manager. Your job is to read client inquiries and draft a short, welcoming email. Keep the tone professional but warm.”

3. Injecting the Data Pills

• The Concept: This is where the magic happens. We write a static prompt, but we inject our dynamic data pills directly into the text box.

• The Formula: We type: "Please write an email to " + [First Name Pill]. "Here is their specific problem: " + [Message Pill].

• The Result: When the workflow runs, the AI reads the specific details of that exact client and generates a brand new data pill containing the custom-written email.

The Hook & Intro

“Welcome back! In Lesson 1, we set up our trigger and caught a test lead. We have a box sitting there with John Doe’s name, email, and his specific business problem. But right now, it’s just raw text. It’s time to bring in the heavy artillery.

It is time to give our workflow a brain. In Step 2, we are going to connect an AI engine—like Google Gemini or ChatGPT—and teach it exactly how to analyze John’s problem and draft a custom reply.”

Point 1: Creating the Persona (The System Prompt)

“Inside your platform, you add Step 2 and select your AI tool. The first thing you will see is a box called the ‘System Prompt.’

(Visual: The screen zooms in on a UI text box labeled ‘System Prompt’. The instructor points to it.)

This is where you give the AI its identity. If you leave this blank, the AI will sound like a generic robot. We don’t want that. We want it to sound like you. So, we type our static instructions: ‘You are an elite customer success manager. Your goal is to draft a welcoming, professional email to a new lead. Keep it under three paragraphs, and always thank them for reaching out.’ You are setting the guardrails.”

Point 2: Injecting the Data Pills

“Now comes the most important part of the entire masterclass. We have to feed the AI the actual lead. We move down to the ‘User Prompt’ box.

We type our static text: ‘Please write a custom email to ‘… and then, we stop typing. We go to our data menu, grab the [First Name] data pill from Step 1, and drop it right into the text box! Then we type, ‘Here is the problem they need help with: ‘ and we drop in the [Message] data pill.

Do you see what just happened? We built a flexible template. The AI isn’t just writing a generic welcome email—it is reading John’s specific marketing problem and addressing it directly.”

The Takeaway & Outro

“Once you set that up, you hit ‘Test Step.’ The AI takes your pills, reads the data, thinks for a few seconds, and spits out a brand new, highly personalized email.

And just like that, the AI has generated a brand new data pill containing our perfect email. But an email isn’t doing us any good just sitting inside Make or Zapier. We have to actually send it! In Lesson 3, we are going to map this AI output into Gmail and hit send. Let’s finish the job. I’ll see you there!”