How to Create a Personalized Sales Coach Using Gemini

If you run a business or lead a sales team, you know that consistent practice is the difference between a closed-won deal and a lost opportunity. However, roleplaying with your reps is time-consuming, and practicing alone doesn’t prepare them for the unexpected objections of a real prospect.

Imagine having a tireless, on-demand sales coach that knows your product inside and out, simulates your toughest buyers, and evaluates your reps against your official company scripts.

By combining your internal company documents with Google NotebookLM and Google Gemini, you can build exactly that. Here is your step-by-step guide to creating a custom AI sales script coach designed to pressure-test your pitches before you ever get on a live call.

Phase 1: Compile Your Sales Playbook

Generative AI is only as effective as the context you provide. To ensure your AI coach evaluates your reps accurately, you need to supply it with your company’s core knowledge.

Gather the following internal documents and save them as PDFs or Google Docs in a single folder:

• The Core Sales Script: Your standard pitch, opening hooks, and discovery questions.
• The Objection Handling Matrix: Approved responses to common pushbacks like “It’s too expensive,” “Send me an email,” or “We already use a competitor.”
• Buyer Personas: Detailed profiles of your Ideal Customer Profiles (ICPs) so the AI knows exactly who it is roleplaying.
• Product Specs & Pricing: The factual limits of your product, deliverables, and pricing tiers so the AI knows what can and cannot be promised.
• FAQs: The most common questions prospects ask on discovery calls.

Pro Tip: Keep these documents updated. Whenever your messaging pivots, be sure to upload the latest versions to your AI coach.

Phase 2: Create Your Source Notebook in NotebookLM

Now that your playbook is assembled, it is time to organize it into a single, cohesive “brain.” For this, we will use Google NotebookLM.

Why use NotebookLM for this step? Because it is specifically designed to ground information securely and accurately. It acts as the perfect centralized hub to hold all of your PDFs and Google Docs without losing the context of your specific business logic.

Here is how to set up your dedicated sales training notebook:

1. Start a New Project: Navigate to NotebookLM and click New Notebook. Name it something clear, like “Sales Script Coach - [Your Company Name].”
2. Add Your Sources: A menu will pop up asking you to add sources. Upload the PDFs, link the Google Docs, or paste the text of the documents you gathered in Phase 1.
3. Let It Index: Give NotebookLM a few moments to read and index your playbook. You will see it generate a quick summary and some suggested questions.

Once your sources are loaded and selected in the sidebar, your custom playbook is officially built. Now, it is time to bring it to life.

Phase 3: Import to Gemini & Set the Prompt

While NotebookLM is perfect for organizing your documents, we want to run our actual interactive coaching sessions directly out of Google Gemini, where the conversational AI truly shines.

Here is how to connect your playbook and instruct the AI:

1. Open Gemini: Navigate to Google Gemini and start a new chat.
2. Import Your Playbook: Click the + (plus icon) next to the chat bar. Select the option to import your newly created NotebookLM playbook so Gemini has full access to your company’s documents.
3. Send the Ultimate Prompt: If you simply tell an AI to “act like a customer,” the roleplay will usually fall flat. It might break character, write your responses for you, or fail to provide actionable feedback. You need to dictate exactly how it should grade you based on your uploaded documents.

Copy and paste this engineered prompt into your Gemini chat:

“Act as an expert Sales Coach and a realistic Prospective Customer. I have imported our company’s official sales playbooks, scripts, FAQs, and product specs from NotebookLM.

Your job is to run a simulation with me to practice my sales calls.

Here are the rules for our simulation: 1. Absorb the Context: Use the imported documents to understand our product, our ideal buyer, and our standard objection-handling frameworks. 2. Play the Customer: Adopt the persona of a qualified but skeptical lead based on the buyer personas provided. Ask tough questions and push back on pricing or features just like a real prospect would. 3. Go Turn-by-Turn: We will speak one turn at a time. Do not write my dialogue for me. Wait for my response before replying. 4. Provide Real-Time Coaching: After every response I give, provide two things in your reply:

• [The Coach]: A brief, direct critique of my response. Did I follow the uploaded script? Did I handle the objection correctly based on our guidelines? How could I improve?
• [The Customer]: Your in-character response to what I just said to keep the roleplay moving.

If you understand, start the simulation by acting as the prospect answering my cold call with: ‘Hello, who is this?’”

Phase 4: Initiate the Live Agent & Start Pitching

Once you hit enter on the prompt above, Gemini will process your instructions and immediately step into character. It will kick off the simulation with the exact phrase you requested: “Hello, who is this?”

Now, the floor is yours.

1. Deliver Your Hook: Type out your actual opening line exactly as you would on a live phone call or Zoom meeting.
2. Review and Adapt: After you hit send, Gemini will reply with its dual-persona format. Read the [The Coach] feedback to see how your response stacked up against your uploaded scripts, then formulate your next reply directly answering [The Customer]‘s dialogue.
3. Go Live (Optional Pro-Move): If you really want to put your reps to the test, have them open the Gemini mobile app on their phone and use the Live (voice) feature to actually speak their responses. This creates a hyper-realistic, real-time phone call environment that will rapidly accelerate their comfort on the phones.

Happy pitching!