Clase 02 - Construye tu primer agente de IA con n8n en solo una clase

Summary

The class teaches how to plan and build an AI-agent workflow with n8n, then demonstrates how to turn the workflow’s data into a dashboard using Lovable.

Key concepts

• AI agents differ from chat assistants: An assistant responds to prompts, while an agent can receive a goal, decide on steps, use connected tools, and carry out a task from start to finish.
• Start with the process, not the software: Before automating, map the routine, identify its steps, and decide which tasks are suitable for automation. If you don’t own the process, investigate how it works first.
• Example business case: Automating routine student-support email replies. The instructor maps the process as: receive an email, understand the request, categorize it, draft a reply, and send it.

n8n tutorial: Email-response agent

The instructor builds the workflow on n8n’s visual canvas using connected nodes:

1. Receive an email: Add a Gmail trigger and connect a Gmail account.
2. Select the needed fields: Use an Edit node to extract the email body and sender address. The instructor presents this as a good practice: pass only relevant data to later steps.
3. Add an AI agent: Connect a chat model—in the demonstration, Gemini—and pass the extracted email data to it.
4. Provide company context: Add a system message with the organization’s information and response guidance. The instructor emphasizes that an unconfigured model tends to give generic answers; context and instructions make its replies more appropriate.
5. Send the reply: Connect Gmail as an agent tool. The demonstration configures the message fields and shows that the model can determine the parameters automatically.
6. Log results: Connect Google Sheets and add a row for each email. The example records the date, sender, category, summary, response, and confidence.
7. Test and publish: Run the workflow with test emails, check the results, and publish only after verifying that it behaves as expected. The instructor recommends reviewing confidence and response quality rather than assuming every AI-generated answer is correct.

Lovable demonstration and dashboard design

Lovable is shown generating a coffee-shop website from a natural-language prompt, including sections such as “Who we are,” prices, and location. The instructor notes that generated pages can be edited and that users should specify the intended audience, relevant information, and problem the system is meant to solve.

The class then demonstrates a dashboard concept for the email workflow. It organizes recorded data into summaries, categories, confidence measures, and visualizations so staff can spot patterns and make decisions without scanning a large spreadsheet.

The broader lesson is to design an end-to-end solution—not just an automation. The presenter frames this as combining the workflow, data storage, and a user-friendly interface.

Other examples and analysis

The instructor shows or describes additional applications, including expense tracking, appointment scheduling, WhatsApp-based customer service, and multi-agent workflows. These are presented as examples of what could be built with more advanced skills.

The class also argues that AI tools can shorten the time required to create prototypes, but stresses that professional systems still require sound process analysis, testing, relevant domain knowledge, and appropriate tool selection. The presenter cites market-demand figures and a Daxus project-cost example; these are claims made during the presentation, not independently verified here.

Reviews, guides, and tutorials

• Tutorial: Build an n8n Gmail-to-AI-agent workflow, send a customized reply, and record the result in Google Sheets.
• Demonstration: Generate a basic website with Lovable and explore a dashboard for workflow metrics.
• Student feedback shared by the presenters: Several comments praised the previous class’s explanations and methodology, and described the material as useful for professional development and business optimization. These testimonials were read aloud; they are not an independent product review.

The video also contains an extended promotion for Daxus training, including courses, certificates, community access, and an enrollment offer. Those details are promotional statements from the presenters.

Main speakers and sources

• Ingrid: Main instructor and presenter.
• Sai (“Professor Sai”): Co-host who reads or responds to chat comments.
• Students and viewers: Their testimonials, questions, and reactions are relayed through the live chat.
• Daxus Latam: The organization and source of the course and business examples presented.