Guide: Getting Started with Gemini 3

A full SOP inside

In partnership with

From Our Sponsor:

Unify Your Coaching Experience to Unlock Revenue Growth

Your clients expect more than content, they expect an experience.

Join 7-figure business coach and former IBM marketing executive, Julie Ciardi, for a free workshop on how she scaled her coaching business to $3.M by prioritizing a premium client experience on Kajabi.

Julie will show how combining courses, coaching, community, and her private podcast into one branded hub creates a cohesive experience that boosts engagement, increases loyalty, and drives long-term revenue, all powered by Kajabi’s full suite.

Guide: Getting Started with Gemini 3

Understanding Gemini 3's New Approach

Google's Gemini 3 represents a fundamental shift in how AI models operate. Unlike previous generations that primarily generate text, Gemini 3 is a reasoning-first model designed to systematically break down tasks, plan workflows, and execute complex projects rather than simply respond to prompts.

What does it mean for us? It means the model can handle strategic planning, multi-step execution, and sophisticated analysis that previously required multiple tools or extensive prompt engineering. Don’t take my word for it, check out the full SOP where I share 3 specific use cases. It’s really, really good.

What's New in Gemini 3

Reasoning-First Architecture

Gemini 3 approaches tasks by planning workflows systematically rather than generating immediate responses. When you request a marketing campaign, the model first decomposes the objective into tactical components, validates required information, and creates a structured execution plan before producing output.

Advanced Reasoning Modes

The model includes Deep Think and advanced thinking modes specifically designed for complex strategic work, detailed analysis, and multi-step planning. These modes provide enhanced reasoning capabilities for projects requiring sophisticated problem-solving.

Enhanced Multimodal Capabilities

Gemini 3 processes documents, slides, images, videos, code, and spreadsheets as equal-class inputs. Upload a product video, brand guidelines PDF, and sales data CSV simultaneously, and the model synthesizes across all modalities to produce integrated outputs including analyses, applications, and user interfaces.

Agentic Workflow Design

The model is built to drive tools and execute automated workflows. It handles complex task sequences, manages multiple components, and produces structured outputs that integrate with existing marketing systems.

How Prompting is Different

Directness Over Verbosity

Old Approach (GPT-style): "I need you to act as an expert marketing strategist with 20 years of experience. Please carefully consider all aspects of my brand positioning and thoroughly analyze..."

Gemini 3 Approach: "Analyze my brand positioning and create a marketing strategy. My brand: [details]. Target audience: [details]."

Gemini 3 prefers concise, action-oriented instructions. Verbose prompt engineering actually reduces performance.

Sequential Reasoning Instructions

Transform long paragraphs into explicit steps using "do X, then Y, then Z" format. Instead of explaining everything at once, break requests into numbered sequences that guide the model's reasoning process.

Constraint Placement Matters

Critical Rule: Place behavioral constraints, role definitions, and brand guidelines at the very beginning of your prompt, not at the end. Gemini 3's reasoning process is anchored by early instructions. Constraints added later may be ignored or overridden.

Reduced Default Verbosity

Gemini 3 provides more concise, efficient answers by default. If you need detailed explanations or a conversational tone, explicitly request it: "Provide a comprehensive explanation with examples" or "Use a friendly, chatty tone."

Long Context Handling

When working with extensive background materials (product catalogs, customer data, research reports):

  1. Place all context first

  2. End with your specific instruction

  3. Use bridging phrases: "Based on the information above..."

This structure allows the model to synthesize information before executing your request.

Core Prompting Techniques

Explicit Planning Instructions

For complex projects, include planning steps in your prompt:

Before creating the campaign:
1. Parse the marketing objective into distinct tactical components
2. Verify that brand positioning and target audience information is complete - if not, ask for clarification
3. Identify proven frameworks that apply to this challenge
4. Create a structured execution plan
5. Validate your understanding before proceeding

Self-Critique for Quality Control

Add validation steps to ensure outputs meet your specific needs:

Before returning your final deliverable, review it against these criteria:
1. Does this address my specific business objective, not generic best practices?
2. Is the tone authentic to my brand voice?
3. Have I flagged any assumptions I made?
4. Does this align with the stated conversion goals?

Progress Tracking for Multi-Step Projects

Request TODO tracking for complex workflows:

Create a TODO list to track this project:
- [ ] Audience research
- [ ] Competitive analysis  
- [ ] Messaging framework
- [ ] Asset creation
- [ ] Performance tracking setup

The model will check off items as it completes them, ensuring nothing gets missed.

This is a long guide, so if you want the full thing, simply head to this Notion page. There you can find the 3 specific use cases I tested.

Do You Love The AI For Ecommerce Sellers Newsletter?

You can help us!

Spread the word to your colleagues or friends who you think would benefit from our weekly insights 🙂 Simply forward this issue.

In addition, we are open to sponsorships. We have more than 55,000 subscribers with 75% of our readers based in the US. To get our rate card and more info, email us at [email protected]

The Quick Read:

The Tools List:

📹 MeetGeek: Automagically record videos, transcribe, summarize, and share insights from every meeting to any tool.

👨‍💼 Aesop - Your AI-powered SaaS to find smart, distinctive names and the stories they tell – for your new brand, business, product, or project (link)*

🖼️ Lummi - Free stock photos made and curated by AI artists.

🤖 Zupyak: Do SEO research and create high-ranking content with AI, without using expensive agencies and contractors. Used by 400,000+ professionals.

🖊️ Intent by Upflowy - Turn your leads' behaviour into AI Summaries.

🗣️ izTalk - Overcome language barriers with instant AI translation

About The Writer:

Jo Lambadjieva is an entrepreneur and AI expert in the e-commerce industry. She is the founder and CEO of Amazing Wave, an agency specializing in AI-driven solutions for e-commerce businesses. With over 13 years of experience in digital marketing, agency work, and e-commerce, Joanna has established herself as a thought leader in integrating AI technologies for business growth.

For Team and Agency AI training book an intro call here.

What did you think of today’s email?