Introduction to Gemma and N8N Automations

Overview of New Automations

• The speaker introduces new one-click automations using Gemma and N8N, highlighting their potential for streamlining various workflows.

• A screenshot is referenced, showcasing an AI-powered Ragn workflow specifically designed for stock earnings.

Features of Gemma 3

• The speaker explains that Gemma 3 can automate processes effectively and mentions a new module in the AI profit boardroom that details its functionality.

• It is noted that Gemma 3 has recently been released and is outperforming other models like Deep Sea version 3 and Lama 3.

Integrating Gemma with N8N

Setting Up Workflows

• The integration of Gemma 3 into N8N workflows is discussed, emphasizing the advantage of using a free API compared to paid alternatives like Claude or Deep Sea.

• Hosting N8N locally allows users to utilize these tools without incurring costs, making it accessible for more users.

Step-by-Step Setup Process

• Instructions are provided on setting up a new workflow in N8N, starting with adding a chat trigger to initiate the AI agent when messages are received.

• Users are guided on how to input Google credentials necessary for connecting with the API through aisstudio.google.com.

Testing and Troubleshooting

Initial Testing Phase

• After setting up the chatbot, users can test its functionality by sending messages; however, initial attempts may encounter errors.

• The speaker emphasizes checking configurations if issues arise during testing, particularly regarding which model (Gemma vs. Gemini) should be used.

Final Adjustments

• It’s clarified that while Gemma may not work as expected for certain applications, switching to Gemini can resolve these issues effectively.

How to Build an AI Agent with SEO Optimization

Setting Up the Content Creation Tool

• The speaker introduces an accessory for building content, indicating a hands-on approach to creating SEO-optimized articles.

• They demonstrate how to plug in a keyword (e.g., "SEO") into an agent designed for generating SEO-optimized articles.

• The process involves taking user input for keywords and outlining the necessary details for article creation, emphasizing customization based on user needs.

Testing the AI Output

• The speaker tests the setup by using "SEO training, Japan" as a keyword and confirms that the content is generated successfully.

• They note that the output is well-formatted and personalized, showcasing effective use of AI in content generation.

Troubleshooting and Model Selection

• An attempt is made to switch back to another model (Gemma), but it results in a bad request error, prompting further troubleshooting.

• The speaker explores using Google's new model via OpenRouter instead of directly accessing Google’s API due to previous issues.

Utilizing OpenRouter for API Access

• After switching models, they confirm successful operation with Gemma 3 through OpenRouter, highlighting its reliability compared to direct Google access.

• Instructions are provided on obtaining an API key from OpenRouter's settings, allowing users flexibility in choosing different models.

Customizing AI Agents with Telegram Integration

• The discussion shifts towards integrating AI agents with Telegram, focusing on customizing them using free models like Gemma from OpenRouter instead of paid options like OpenAI.

• A demonstration follows where they set up an AI agent that listens for incoming messages on Telegram and responds accordingly.

Final Testing and Observations

• Successful testing shows that the integration works smoothly; responses from both Google Gemma 3 and OpenAI are compared for performance evaluation.

Understanding the Workflow of AI Agents

Overview of Content and Personality

• The content is described as being more engaging and friendly, resembling a conversational tone rather than a straightforward AI response.

• The speaker compares different AI models, noting that while Gemma 3 has personality, it still feels distinctly like an AI compared to other options.

Setting Up Workflows

• The speaker demonstrates how to set up workflows step by step, emphasizing ease of use with the tools available.

• Users can access various agents within the platform, showcasing a range of functionalities tailored for specific tasks.

Importing and Utilizing Agents

• Instructions are provided on downloading JSON files for deep seek agents and importing them into the system seamlessly.

• Switching between APIs is highlighted as simple; users can easily integrate Gemma into their workflows with minimal clicks.

Research Agent Functionality

• A research agent is introduced that utilizes an open router chat model to gather information based on user input (e.g., Google Gemma API).

• The process involves multiple steps where the agent compiles data before moving onto writing tasks such as introductions or emails.

Human-in-the-loop Approval Process

• An important feature discussed is the human approval mechanism integrated into workflows, allowing users to review content before proceeding further. This ensures quality control in generated outputs.

• If content requires revision after approval, it loops back through designated agents for rewriting until satisfactory results are achieved. This iterative process enhances output quality significantly.

Conclusion on Workflow Capabilities

How to Set Up AI Agents and Scrape the Internet

Setting Up Human in the Loop Workflows

• Discusses the importance of human oversight in content approval processes, where content can be approved or declined before moving to the next workflow stage.

• Mentions using OpenRooter Gemma 3 for this process, emphasizing its efficiency with a one-click setup via the Gemma API.

Scraping the Internet with AI Agents

• Introduces a powerful script that allows users to scrape web pages using AI agents, specifically mentioning HTTP requests.

• Demonstrates how to input a URL into Gemma 3 to scrape any desired webpage and highlights customization options for connecting it to chat functionalities.

Generating Content Ideas from Web Pages

• Explains how an AI agent can analyze scraped data from web pages and generate blog ideas based on that information.

• Emphasizes flexibility in scraping methods, including HTML extraction and website redesigning for research purposes.

Troubleshooting API Issues

• Shares insights on common errors encountered during live demonstrations and stresses the importance of real-time troubleshooting skills.

• Suggests switching between different APIs (e.g., from Gemma 3 to Claw 3.7) when issues arise, noting personal preference for Claw 3.7 due to its effectiveness in generating human-like content.

Iterative Testing and Community Support

• Highlights the necessity of patience while working with new technologies like AI agents, as initial setups may not work perfectly.

• Encourages joining community platforms like the AI profit boardroom for collaborative problem-solving and support among members facing similar challenges.

Quick Setup of AI Agents

• Describes an efficient method for importing pre-configured workflows into new projects within the N A 10 agent section.

Community Goals and Automation Tools

Purpose of the Community

• The community aims to save members hundreds of hours rather than solely focusing on business growth with AI. This approach is based on their success in scaling to over $300,000 a month.

• Members can post any questions they have, fostering an environment of support and collaboration.

Automation Offerings

• The community provides access to various automation tools including email content generation, video meeting AI agents, web scraping, SEO automation, and crash courses. New ideas for courses are continuously being developed.

• Upcoming courses include topics like "Go High Level Automations" and business optimizations that aim to enhance operational efficiency.

Downloadable Resources

Accessing Automation Files

• Users can download various AI agents such as a Telegram assistant, video transcript assistant, SEO content writer, and more directly from the community platform. These files can be easily imported into NA10 systems using a free API from Gemma.

Free Strategy Sessions

• A free one-on-one SEO strategy session is available where participants can learn how to grow websites significantly and generate substantial sales through effective link-building strategies. Questions can be asked live during these sessions for personalized assistance.

Innovative AI Agent Demonstration

Inbound Call Automation

• An exciting feature demonstrated is an AI agent capable of answering inbound phone calls linked to a website's phone number, automating the entire process by taking messages and providing transcripts later for team access.

Live Interaction Example

• A live demonstration shows how the AI agent interacts with callers by gathering information needed for booking SEO strategy sessions while maintaining a conversational tone throughout the interaction. This showcases its ability to handle inquiries effectively without human intervention.

Benefits of Using AI Agents

Advantages Over Human Operators

AI-Powered Solutions for Business Efficiency

The Role of AI in Replacing Traditional Roles

• Businesses can leverage AI agents to eliminate the need for managing human roles, such as receptionists, thereby streamlining operations and reducing costs.

• Transitioning from a human receptionist to an AI agent could save thousands of dollars annually and significantly reduce hours spent on routine tasks.

Setting Up an AI Agent

• A tutorial is available within the "AI Profit Border" that guides users through setting up an AI-powered inbound call agent using 11 Labs technology.

• Users can track performance metrics through a dashboard that displays inbound calls, average call duration, and associated costs.

Enhancing Customer Interaction with AI

• The speaker has created a voice clone that answers calls, allowing for personalized interactions while saving time.

• An embedded "start call" button on websites connects potential clients directly to the AI agent after accepting terms and conditions.

Demonstrating the Functionality of the AI Agent

• A live demonstration shows how callers interact with the AI agent, which can engage in meaningful conversations about topics like SEO and business strategies.

• The AI agent not only provides information but also collects client details during calls, enhancing lead conversion rates compared to traditional contact forms.

Steps to Implement Your Own AI Agent

How to Create an AI Receptionist

Setting Up Your AI Agent

• When establishing a personal brand or as a CEO, it's crucial to have a real person answer calls rather than relying solely on clones. Training the AI can be done with just 10 minutes of audio footage.

• The process begins by creating an agent using a blank template, which can be named (e.g., "Julian Goldie Receptionist"). It's essential to set up all relevant details and add a knowledge base for effective responses.

• Linking the AI agent to Twilio allows it to connect with phone numbers. Additionally, training data from sales pages and websites enhances the agent's knowledge about services offered.

Enhancing AI Capabilities

• Users can scrape live content from their website or specific pages instead of uploading PDFs, making it easier to keep the AI updated with current information.

• Adjusting the "temperature" setting influences how creative or random the responses are. Options include using various models like Gemini or GPT 4.0 for answering queries effectively.

Customizing User Interaction

• System prompts define how the AI interacts; for instance, it can be programmed as a friendly support agent who takes messages and ensures follow-up communication.

• Collecting client information such as phone numbers and email addresses is vital during interactions. This setup aims at generating efficient responses while maintaining user engagement.

Addressing User Needs

• The speaker encourages audience interaction by inviting questions about AI usage, emphasizing that feedback is important for ensuring helpfulness in provided solutions.

• Highlighting cost-effectiveness, having an AI receptionist is more engaging than traditional virtual offices that operate only during business hours, providing 24/7 availability.

Language and Security Features

• The system supports multiple languages for conversations, allowing customization based on user preferences. This feature broadens accessibility for diverse clientele.

User Interface Customization

Changing Visual Elements

• The background color and text color of the interface can be customized for a cleaner look, with white being preferred for its aesthetic appeal.

• After refreshing the website, changes such as an orange border color are visible, indicating successful updates to the UI.

Modifying Widget Text

• The text displayed in widgets can be modified; for instance, changing it to "Speak to an agent" enhances clarity and user engagement.

• Upon refreshing the site, users can see updated call-to-action (CTA) phrases like "Need help?" which improve interaction.

AI Agent Functionality

Consent and Language Options

• Users must consent to recording their communications with third-party service providers as per privacy policies. This is crucial for transparency.

• Language selection options are available, allowing users to choose their preferred language for better communication.

Call Management Features

• Phone numbers can be imported into the system for inbound calls, linking them directly to specific agents for efficient handling.

• Call history features provide detailed insights into call duration, success rates, and timestamps of interactions.

Performance Tracking and Automation Benefits

Efficiency Gains

• The AI system is described as one of the most powerful tools seen in automating tasks that traditionally consume significant time.

• Automatic tracking through integrated systems reduces administrative burdens on staff, streamlining operations significantly.

Community Support and Resources

• A tutorial is available within the platform detailing how to set up an AI-powered inbound call agent linked with Twilio services.

• Users have access to a community of 455 members where they can seek advice or assistance regarding automation processes.

Additional Offerings and Discounts

Yearly Savings Opportunities

• A yearly subscription offers approximately 30% savings compared to monthly plans, making it financially advantageous for long-term users.

Personalized Sessions

AI Tools for Structural Engineering and Content Creation

Introduction to AI Tools in Engineering

• Kinenex, a structural engineer from South Africa, discusses the integration of AI tools into engineering practices to enhance report generation and overall efficiency.

• The speaker highlights the availability of over 50 SEO tools developed within the AI Success Lab, emphasizing their effectiveness in content creation and keyword research.

Automation and Efficiency

• Emphasizes the importance of automation in freeing up time for engineers, allowing them to focus on scaling their work by eliminating or delegating tasks.

• Introduces a comparison between various AI models: Gemma 3, Grock, ChatGPT, and Claude, aiming to determine which tool performs best.

Overview of Gemma 3

• Provides an introduction to Gemma 3 as Google's latest model capable of running on a single GPU with impressive benchmarks against other models.

• Mentions that users can access Gemma 3 for free at aistudio.google.com.

Testing AI Models

• Begins testing different prompts across the selected AI models to evaluate reasoning capabilities.

• Sets up a reasoning challenge involving picking an apple from a tree during winter to assess logical thinking among the models.

Evaluation of Responses

• Notes that Gemma 3 provides detailed answers with multiple solutions while demonstrating logical reasoning about seasonal fruit availability.

• Compares responses from Grock and Claude; while Grock offers fewer ideas, Claude quickly identifies that apples do not grow in winter.

Conclusion of Model Comparison

• Concludes that Gemma 3 provided the most practical solutions followed by Claude. ChatGPT did not perform well in this instance.

AI Content Creation Comparison

Preference for AI Tools

• The speaker expresses a personal preference for using Claude, noting it as the best tool for creating content. They mention generating an SEO training article with 1000 words.

Performance of Different AI Models

• Gemma is noted to be slower in generating content but produces nicely formatted text. The speaker highlights the competitive nature of SEO in London and how established brands impact visibility.

• A comparison is made between Gemma's writing style and Claude's, indicating that Gemma's output feels less humanized and more like typical AI-generated text.

• ChatGPT’s content is described as better than Gemma’s but still not matching Claude’s quality. The speaker emphasizes the importance of human-like writing in SEO content.

Humanization and Detectability

• Groc's output is recognized for its keyword optimization, though it doesn't quite reach Claude's level of humanization. The ranking established: Claude first, Groc second, ChatGPT third, and Gemma last.

• The speaker plans to assess how humanized each model's content appears using Zero GPT to measure detectability.

• Claude scores impressively low at 0.9% detectable as AI-generated text, while Groc comes in at 15%, ChatGPT at 19%, and Gemma significantly higher at 39%.

Summary of Findings

• Overall rankings based on humanization are reiterated: Claude leads with the most natural writing style followed by Groc, then ChatGPT, with Gemma lagging behind due to its perceived "AI fluff."

• The speaker notes that while performance may vary when running locally versus in a studio setting, Claude consistently outperforms others regarding speed and quality of output.

Coding Capabilities Assessment

• Transitioning to coding tasks, the speaker intends to test each model by asking them to create a self-playing snake game using Python with HTML GUI specifications.

• While expressing dissatisfaction with Groc compared to Claude overall, they acknowledge Groc’s quick response time during coding tasks.

Comparison of AI Models: Claude, Gemma, and Others

Initial Testing of AI Models

• The speaker tests the functionality of various AI models, noting that Claude is performing well in terms of controls and user interface.

• Attempts to preview Gemma directly within the chat are unsuccessful; HTML output is copied for further testing.

• A demonstration using Liveweave shows that while some features work, others fail repeatedly.

Performance Evaluation

• In a comparative analysis, Gemma ranks last due to its lack of functionality in play mode; Claude leads with superior UI and design features.

• Grog ranks second for speed, while ChatGPT comes in third; Gemma excels only in reasoning tasks compared to other models.

Local Running Capabilities

• Gemma's ability to run locally is highlighted as a significant advantage over Claude and Grog, which cannot be run offline.

• Instructions are provided on how to set up Gemma 3 locally using Alarm 3 software.

API Integration and Workflow Creation

• The speaker discusses building AI agents using OpenRouter with Gemma 3 as an example for creating workflows in n8n.

• Templates for AI agents are available within the community resources, allowing users to import them into their workflows easily.

Conclusion on Model Usage

• While Gemma 3 is effective for building free AI agents and offers a free API through OpenRouter, it is not recommended for coding or content creation tasks.