How I Automated an SEO Agency with 15 AI Agents (No-Code)

Summary of YouTube Video: Automating Content Creation for SEO Agencies with AI Agents

Key Concepts:

AI Agents: Software entities designed to perform specific tasks autonomously.
Multi-Agent Systems: A network of interacting AI agents working together to achieve a common goal.
SEO (Search Engine Optimization): Optimizing online content to rank higher in search engine results.
Content Brief: A document outlining the requirements and guidelines for creating a specific piece of content.
Relevance AI: A no-code platform for building AI agents and multi-agent systems.
Make.com: An automation platform used to integrate different software applications.
Webhooks: Automated HTTP requests triggered by specific events in one application to update another.
Vector Data Search: A method of searching for data based on similarity in a high-dimensional space.

1. Introduction

The video demonstrates how to automate the entire content creation process for an SEO agency specializing in e-commerce brands using a team of 15 AI agents. The system handles in-depth research, content brief creation, article writing in the brand's tone of voice, product linking, and publishing to the client's website with FAQs, tables, and images. It also supports the creation of product review pages, product description pages, category pages, and landing pages.

2. Client Context and Results

The system was implemented for an SEO agency serving fashion e-commerce brands. The presenter waited to showcase the system until there was proof of its effectiveness in ranking content on Google. The agency provided screenshots of results from three clients:

Client 1: Clicks increased from approximately 125 to 300 per day over three months.
Client 2: Similar growth pattern, increasing from around 125 to 300 clicks per day.
Client 3: A larger client saw a steeper increase from 750 to nearly 2,000 clicks per day.

The agency attributed the biggest impact to the increased content output volume. Traditionally, they could produce 150 pieces of content every two and a half months at $50 per piece. With the AI system, they can produce up to 150 pieces of content per day at $1.50 per piece.

3. System Architecture and Workflow

The system is built using Relevance AI and Make.com. Relevance AI is used for building the multi-agent system, while Make.com facilitates integrations with third-party platforms like Webflow and WordPress for content publishing.

The system operates with a three-layered agent structure:

Director (SEO Director): Orchestrates the entire process, updates the CRM, and publishes content to the client's website.
Managers (e.g., Research Manager, Content Brief Manager): Delegates tasks to sub-agents and evaluates their work.
Sub-Agents (e.g., Keyword Agent, Topic Research Agent, Blog Post Writer): Perform specific tasks such as keyword research, topic research, competitor analysis, and content writing.

The workflow is triggered from a CRM (replicated in AirTable for the demo). The client inputs the URL, topic, content type, language, keywords (optional), competitors (optional), and any additional notes.

4. Detailed Process Breakdown

Trigger: The process starts when the client fills out the required information in the CRM (AirTable).
SEO Director: Receives the information from the CRM and orchestrates the workflow.
Research Manager:
- Instructed by the SEO Director.
- Delegates tasks to three sub-agents:
  - Keyword Agent: (Not included in the demo due to SEMrush requirement) Uses SEMrush to find the best ROI keywords and top-ranked articles.
  - Topic Research Agent: Uses Google Search, web scraping, and (ideally) Perplexity to conduct in-depth research on the topic and keywords.
  - Competitor Research Agent: Analyzes top-ranked articles to identify opportunities for outranking competitors.
- Evaluates the reports from the sub-agents for comprehensiveness.
Content Brief Manager:
- Creates a content brief based on the research analysis.
- Delegates the task to one of five sub-agents, depending on the content type (blog post, category page, landing page, product review, or product description).
- In the client's system, the content brief is sent back to the SEO director for approval before proceeding.
Content Creation:
- Content Drafter: Creates an initial draft of the content, using a database of blog examples for different categories and clients to understand the tone of voice and brand.
- Content Writer (Specific to Content Type): Finalizes the content piece.
  - Equipped with a product database to pull product information, images, and links.
  - Can retrieve brand identity from the website.
  - Can use stock photos if necessary.
- Backlink Agent: (Not included in the demo due to SEMrush requirement) Adds external backlinks to the article using SEMrush data.
SEO Director: Posts the finalized content to the client's website and updates the CRM.

5. Demo and Examples

The presenter demonstrates the system by generating three types of content for nike.com:

Blog Post: "Top 5 Best Nike Running Shoes for Outdoor Enthusiasts." The system pulls product images and key features from the product database.
Blog Post: "Mastering Your Outdoor Running Technique: Tips for Every Runner." The system uses stock photos.
Category Page: A category description page with FAQs.

The demo highlights the system's ability to generate different content types and integrate product information and images.

6. Technical Details and Implementation

Relevance AI: The presenter provides a brief overview of the agent configurations in Relevance AI, including the prompts and tools used by each agent.
Make.com: The presenter explains how to set up a webhook trigger in AirTable to initiate the content creation process in Relevance AI. This involves using an AirTable automation with a "Run Script" action to send the record ID to Make.com. The Make.com scenario then retrieves the record details and triggers the Relevance AI agent. The script code for the AirTable automation is promised to be shared in the presenter's free community.

7. Key Arguments and Perspectives

AI Content Can Rank: The presenter challenges the common objection that AI-generated content cannot rank on Google, providing evidence from the client's results.
Volume is Important for E-commerce SEO: The presenter emphasizes the importance of content volume for driving clicks and traffic in the e-commerce sector.
Multi-Agent Systems Offer Control and Quality: The layered agent structure allows for control agents (managers) to evaluate the work of sub-agents, ensuring quality and comprehensiveness.
Replicating Real-World Workflows: The system is designed to closely mimic the actual workflow of content creation in SEO agencies.

8. Notable Quotes

(Regarding the impact of the system) "The biggest impact this system made is just the output volume."
(Explaining the role of the Research Manager) "This is sort of the advantage to building these multi-layered agent systems we can have sort of these control agents right that can make sure that everything's done what should be done by these uh sub agents of them."

9. Conclusion

The video showcases a sophisticated AI-powered system for automating content creation for SEO agencies. By leveraging multi-agent systems, the solution significantly increases content output volume while maintaining quality and relevance. The system's modular design and integration with various tools and platforms make it a powerful solution for e-commerce SEO. While the presenter doesn't provide the complete template for free, they offer valuable insights into the system's architecture and implementation, along with resources for setting up similar workflows. The key takeaway is that AI, when implemented strategically with a well-designed multi-agent system, can be a game-changer for content creation and SEO.