n8n for Beginners: Google Workspace Automation with AI

In this video, I’ll show you how to use n8n to create a new folder in Google Drive. I’ll start with a simple manual trigger, add the Google Drive node, and configure it to create a folder inside My Drive. Then I’ll make the workflow more dynamic by adding a timestamp to the folder name. Finally, I’ll replace the manual trigger with a form submission, so the folder name can be entered when the workflow runs, and I’ll show how that input is passed into the Google Drive node to create the folder automatically.

In this video, I’ll show you how to use n8n to move files between folders in Google Drive. I’ll start with two folders, Incoming and Processed, then build a workflow that searches the Incoming folder and lists the files inside it. I’ll show you how to use the folder ID from Google Drive, how to return all matching files, and how n8n passes file data, including each file’s unique ID, into the next node.

Then I’ll add a second Google Drive node to move each file into the Processed folder. I’ll show you how to use expressions so the workflow automatically moves every file returned by the first node, instead of selecting files manually. By the end, you’ll have a simple workflow that can organize files on Google Drive by finding them in one folder and moving them into another.

In this video, I’ll show you how to turn the previous file-moving workflow into an automatic Google Drive monitoring workflow. Instead of starting the automation manually, I’ll replace the manual trigger with a Google Drive trigger that watches a specific Incoming folder for newly created files.

I’ll show you how to use the folder ID, set the trigger to watch for new files, and publish the workflow so it runs automatically in the background. Then I’ll test it by creating new Google Docs inside the Incoming folder and showing how n8n detects them, triggers the workflow, and moves them into the Processed folder. I’ll also show you where to view the workflow executions and how to unpublish the workflow when you want to stop it.

In this video, I’ll build on the previous Google Drive workflow and show you how to log file activity into Google Sheets. I’ll start with the automation that monitors an Incoming folder, moves files into a Processed folder, and then add a new step to capture useful information about each file.

I’ll create a Google Sheet called Drive Log with columns for the timestamp, filename, file ID, and file size. Then I’ll use the Edit Fields node in n8n to extract and organize those values from the workflow data. After that, I’ll add a Google Sheets node and configure it to append a new row to the spreadsheet each time a file is processed.

By the end, the workflow will not only move files in Google Drive, but also keep a running log in Google Sheets showing which files were processed and when.

In this video, I’ll show you how to read rows from Google Sheets, use that data to control a workflow, and then update the sheet with the results. I’ll start with a Drive Control spreadsheet that contains file IDs and an action column telling n8n whether each file should be archived or sent for processing.

I’ll use the Google Sheets node to read the rows from the spreadsheet, then add an If node to check whether each row says Archive or Process. Based on that decision, the workflow will move each file to either an Archive folder or an Incoming folder in Google Drive.

Finally, I’ll add Google Sheets update nodes to write information back into the spreadsheet, including the file name and a status message. By the end, you’ll have a workflow where a Google Sheet acts as a simple control panel for managing files in Google Drive.

In this video, I’ll show you how to use two Google Sheets together in one n8n automation. I’ll work with a Drive Control sheet, which lists files and their current status, and a Drive Log sheet, which records files that have actually been processed.

I’ll build a workflow that reads rows from both sheets, filters the Drive Control sheet to find files marked as sent for processing, and then checks whether those file IDs also appear in the Drive Log. To do that, I’ll introduce the Merge node and show how it can match rows from two different data sources using the file ID.

Finally, I’ll update the Drive Control sheet so any file that has been processed is marked as processed, with the correct timestamp copied from the Drive Log. By the end, you’ll see how n8n can compare data across multiple spreadsheets and keep your tracking sheets synchronized automatically.

In this video, I’ll show you how to set up OpenRouter credentials so n8n can connect to external AI models. I’ll briefly explain what OpenRouter is, how its credit system works, and where you can view your activity, usage logs, token counts, and costs.

Then I’ll create a new API key inside OpenRouter and add it as a credential in n8n. To test that everything is working, I’ll build a simple workflow with a manual trigger, an AI Agent node, and an OpenRouter chat model. Finally, I’ll run the workflow, generate a quick joke, and check the OpenRouter logs to confirm that the request was processed and charged correctly.

In this video, I’ll show you how to connect n8n to a local AI model running on your own computer with Ollama. I’ll briefly explain how to download Ollama, install a local model, and make sure it is running before connecting it to n8n.

Then I’ll create an Ollama credential inside n8n and show the important base URL change needed when n8n is running through Docker. After that, I’ll go back to the simple joke workflow from the previous video, replace the OpenRouter chat model with an Ollama model, and run the workflow again.

By the end, you’ll see how n8n can use either a paid external AI service like OpenRouter or a locally installed AI model through Ollama.

Setting up more AI using aIn this video, I’ll show you another way to connect local AI to n8n using LM Studio. I’ll start by showing how LM Studio works, including how to download language models, select an installed model, and chat with it directly on your own computer.

Then I’ll show you the key step for using LM Studio with n8n: enabling the local server from the Developer tab. After that, I’ll create an OpenAI-style credential in n8n, using the local LM Studio server URL so n8n can communicate with the models running on my machine.

Finally, I’ll update the simple AI Agent workflow and connect it to the LM Studio model instead of Ollama or OpenRouter. I’ll test it with a joke prompt and a poem prompt, so you can see how n8n can use LM Studio as another local AI option for your automations. different and very powerful free tool.

In this video, I’ll take a closer look at the AI Agent node in n8n and show how it receives prompts and passes them to an AI model. I’ll start by adding an AI Agent to a workflow and explaining how n8n automatically creates a chat trigger when the agent needs user input.

I’ll show how the “When Chat Message Received” trigger works, how the AI Agent uses that chat message as its prompt, and how to connect a local Ollama chat model so the agent can generate a response. Then I’ll switch to a different setup using a regular trigger and show why, in that case, we need to define the prompt manually inside the AI Agent.

Finally, I’ll demonstrate how to pass data from a form submission into the AI Agent prompt. I’ll show how a form field can become part of the prompt, and how you can combine fixed instructions with dynamic user input to create more flexible AI-powered automations.

In this video, I’ll introduce the Google Docs node in n8n and show how to create, retrieve, and update a Google document. I’ll start by creating a new Google Doc from n8n, then use the document ID to retrieve that same file from Google Drive.

After that, I’ll add an Update Document step and insert simple text into the body of the document. Then I’ll make the workflow more useful by adding an AI Agent, connecting it to a local Ollama model, and using the AI output as the text that gets inserted into the Google Doc.

I’ll also show what happens when AI-generated text includes Markdown, and how to enable Markdown support inside Google Docs. By the end, you’ll have a simple workflow that creates or finds a document, generates text with AI, and writes that content directly into Google Docs.

In this video, I’ll build a practical n8n workflow that uses AI to create article outlines and save them directly into Google Docs. I’ll start with a simple version of the automation using a manual trigger, an AI Agent, and a Google Docs node. The AI Agent will generate an outline from a structured prompt, then n8n will create a new Google Doc and insert the AI-generated outline into it.

I’ll also show how to troubleshoot the AI model connection by switching from LM Studio to Ollama when needed, and how to use the AI Agent output inside the Google Docs Update Document node. Once the basic version is working, I’ll make the workflow more flexible by replacing the manual trigger with a form submission.

The form will ask for the article topic, target audience, and purpose, then pass those values into the AI prompt. I’ll also use the topic to create a dynamic Google Doc title. By the end, you’ll have a reusable workflow where you can enter any topic, audience, and purpose, and n8n will generate a complete article outline in Google Docs automatically.

In this video, I’ll improve the article outline workflow so the generated outlines are cleaner and more useful for the next automation. I’ll start by switching the AI model from Ollama/Gemma to an LM Studio model, so the output no longer includes the extra thinking process inside the Google Doc.

Then I’ll add another Google Docs update step to insert the article topic, audience, and purpose at the top of the document. This means each outline contains the key instructions needed for a future workflow, even if it is used independently from the original form submission.

Finally, I’ll change the Google Docs create step so new outlines are saved inside a dedicated Outlines folder in Google Drive instead of the root folder. By the end, the workflow will generate cleaner, better-organized article outlines that are ready to be used in the next step: turning an outline into a full article.

In this video, I’ll show you how to turn an article outline into a full article using n8n, Google Docs, and AI. I’ll start with a folder of saved outlines in Google Drive, then build a workflow where I enter the document ID of one outline and use the Google Docs node to retrieve its contents.

From there, I’ll add an AI Agent connected to OpenRouter and create a structured prompt that expands the outline into a more detailed version first. Then I’ll use a second AI Agent to turn that expanded outline into a complete SEO-friendly article, including an SEO title, keywords, a meta description, and the full article text.

I’ll also show how to ask the AI to return the result as JSON, then use the Edit Fields node to separate the title, keywords, meta description, and article into individual fields. Finally, I’ll create a new Google Doc in an Articles folder, use the SEO title as the document name, and write the finished article into the document. At the end, I’ll set a challenge to adapt the workflow so it can process every outline in a folder automatically, instead of working with one document ID at a time.

In this video, I’ll go through a solution to the homework challenge from the previous lesson. Instead of entering one outline document ID at a time, I’ll modify the workflow so it can read every outline inside the Outlines folder and turn each one into a full article automatically.

I’ll start by duplicating the original Outline to Article workflow, replacing the form submission trigger with a manual trigger, and adding a Google Drive search step to find all files in the Outlines folder. Then I’ll update the Google Docs step so it retrieves each outline by its file ID and passes the content through the existing AI article-generation workflow.

I’ll show how n8n processes each outline one by one, creates a finished article for each, and saves the results into the Articles folder. I’ll also check the OpenRouter usage cost for generating multiple articles. Finally, I’ll add a cleanup step that moves completed outlines into a Processed folder, so they won’t be picked up again the next time the automation runs.

In this video, I’ll show you how to take the AI-generated article workflow and publish the finished content to a WordPress site instead of only saving it to Google Drive. I’ll start by creating a WordPress API credential in n8n using a WordPress username, an application password, and the website URL.

Then I’ll duplicate the existing outline-to-article workflow and replace the Google Docs saving steps with a WordPress node. I’ll show how to create a new WordPress post, use the AI-generated SEO title as the post title, and convert the article from Markdown into HTML before sending it to WordPress. I’ll also set the post status to draft so it can be reviewed before publishing.

Finally, I’ll add an HTTP Request node to update the post excerpt using the AI-generated meta description. By the end, the workflow will be able to create a complete draft WordPress post from an AI-generated article, including formatted content and an excerpt.

In this video, I’ll show you how to create a Gmail automation in n8n that sends a notification when a new email arrives. I’ll start by setting up a Gmail trigger that polls the inbox and fires when a new message is received.

Then I’ll use an Edit Fields node to pull out key details from the email, including the subject, sender, and snippet, so we can see the information being passed through the workflow.

After that, I’ll introduce Pushinator, a third-party push notification service that can send alerts to a mobile device. I’ll show how to create a Pushinator API token, add it as a credential in n8n, create a notification channel, and connect the mobile app by scanning the QR code.

Finally, I’ll add the Pushinator node to the workflow and build a notification message using the email details from Gmail. By the end, the workflow will be able to detect a new Gmail message and send a push notification straight to your phone.

In this video, I’ll show you how to log new Gmail messages into a Google Sheet using n8n. I’ll start from the previous Gmail notification workflow, duplicate it, and remove the Pushinator step so we can focus on recording email details instead.

I’ll use a Gmail trigger to detect a new email, then use an Edit Fields node to extract the message ID, subject, snippet, and sender. After that, I’ll add a Google Sheets node and configure it to append a new row to a Support Request spreadsheet, mapping each email detail into the correct column.

I’ll also show why using the Google Sheet ID can be more reliable than selecting the sheet by name, and how to refresh the column list if new headers don’t appear in n8n. Finally, I’ll add a Gmail step to mark the email as read after it has been logged, so the inbox stays organized.

In this video, I’ll build on the previous Gmail-to-Sheets workflow and show you how to automatically reply to incoming emails with AI. I’ll duplicate the existing workflow, keep the Gmail trigger and spreadsheet logging steps, and then add an AI Agent to draft a polite response based on the sender, subject, and email snippet.

I’ll connect the AI Agent to OpenRouter using GPT 4.1 mini, then write a simple prompt that tells the AI to act as an email assistant and create a concise reply. After that, I’ll add a Gmail Send node to send the reply back to the original sender, using the original subject line with “RE:” added at the front.

Finally, I’ll update the Support Request spreadsheet with the AI-generated reply and the reply date. By the end, the workflow will detect a new Gmail message, log it in Google Sheets, generate and send an AI-written response, mark the original email as read, and keep a record of the reply in the spreadsheet.

This section moves you from following step-by-step lessons to building your own n8n automations from realistic project briefs. Each brief gives you a goal and enough guidance to plan the workflow yourself, then you can compare your solution with mine afterward. The aim is to help you think like an automation builder by choosing triggers, connecting Google Workspace tools, testing your workflow, and improving it through practice.

This brief asks you to build a Content Tone Rewriter automation that takes original text, sends it to AI with tone or style instructions, and produces a rewritten version. The user should be able to provide the text and choose or describe the rewrite style, such as formal, friendly, shorter, simpler, persuasive, or professional. The finished result should then be saved somewhere useful, such as a Google Doc, Google Sheet, Drive folder, or sent by Gmail.

In this video, I’ll build a Tone Rewriter workflow in n8n that rewrites text in a different style and saves the result to Google Docs. I’ll start by creating a form where I can enter a document title, choose a target tone, select a formality level, choose the preferred length, and paste in the text I want to rewrite.

Then I’ll connect the form to an AI Agent and use a structured prompt that tells the AI how to rewrite the text while preserving the original meaning and facts. I’ll show how the prompt uses the form inputs, including the target tone, formality score, length preference, and original text.

After that, I’ll create a Google Doc in a dedicated Tonal Rewrites folder and write the AI-generated rewrite into the document. I’ll test the workflow by rewriting a deliberately rude customer service email in a professional tone, then again in a friendly tone.

Finally, I’ll show how easy it is to extend the workflow by adding a new tone option, such as “comedian,” and updating the AI prompt so the automation can rewrite the same text in a completely different style.

This brief asks you to build a Grammar Correction Tool that checks written content for spelling, grammar, and language issues using AI. The workflow should receive text from a form or Google Doc, let the user choose the editing level, such as spelling only, full grammar check, or line editing, and specify the language style, such as US English or UK English. The automation should then generate either a corrected version, an editing report with explanations, or both, and save the result somewhere useful.

In this video, I’ll start building a Grammar and Line Editor workflow in n8n. I’ll create a form where I can enter a project title, paste in a Google Docs article ID, choose the level of editing, and select whether the document should be checked in UK or US English.

Then I’ll retrieve the article from Google Docs and use an If node to split the workflow into two paths: one for spelling and grammar checks, and another for line editing. In this lesson, I’ll focus on the spelling and grammar route, using an AI Agent with a detailed proofreading prompt that checks spelling, punctuation, grammar, capitalization, tense, and English variant consistency.

I’ll show how the prompt is built using both fixed instructions and form variables, then save the AI-generated report into a new Google Doc inside a Grammar folder. Finally, I’ll compare the results from a smaller local model with a stronger model through OpenRouter, showing how model choice can make a big difference in proofreading quality.

In this video, I’ll complete the Grammar and Line Editor workflow by building the line editing branch. I’ll connect the Line Editor AI Agent to the same AI model used for the spelling and grammar route, then create a separate prompt focused specifically on line editing.

I’ll show how the line editor prompt asks the AI to review the document line by line, looking for clarity, flow, word choice, repetition, sentence structure, readability, and phrasing, while preserving the original meaning and style. I’ll also use a structured Markdown report format so the AI returns each issue with the reason, original text, and suggested improvement.

Then I’ll create a new Google Doc for the line edit report and save the AI output into the Grammar folder. I’ll test the workflow with Gemini 2.5 Pro through OpenRouter and review the quality of the suggestions, including improvements to word choice and redundancy. Finally, I’ll compare the result with a smaller free model to show how much the choice of AI model affects the quality of editing.

This brief asks you to build an AI Humanizer workflow that reviews written content and identifies patterns that may make it sound AI-generated. The automation should first create or retrieve a stored list of common AI writing tells, then compare a submitted text against that list and flag possible issues such as repetition, vague wording, over-polished structure, or predictable phrasing. The final output should be saved somewhere useful as a feedback report, a rewritten version, or both.

In this video, I’ll show you how I would build an AI Humanizer workflow in n8n. Rather than asking AI to rewrite the whole document automatically, I’ll create a report that highlights common AI writing patterns so I can review the suggestions and edit the original text manually.

I’ll start by researching common AI tells using NotebookLM, then save those tells into a Google Doc so the automation can use them as a reference. After that, I’ll build a form in n8n where I can enter a project title and paste in the text I want to check.

Then I’ll use Google Docs to retrieve the AI tells document, pass both the reference list and the submitted text into an AI Agent, and ask it to identify passages that sound machine-generated. The report will flag issues such as repetitive structures, “not just A but B” phrasing, rule-of-three patterns, overused vocabulary, excessive hedging, and em dashes.

Finally, I’ll save the AI-generated report into a Google Docs folder and test the workflow with both a short example and a longer AI-written story. By the end, you’ll have a workflow that can scan text for AI-style patterns and produce a practical editing report you can use to humanize the writing yourself.

This brief asks you to build a Social Media Content Factory that turns an existing article into platform-specific posts for X, LinkedIn, and Facebook. The automation should retrieve or receive the article, send it to AI, and generate posts that match each platform’s style, including elements such as emojis, hashtags, calls to action, and different tones. It should also create image prompt ideas for the posts and save everything in a clearly organized Google Doc in Google Drive.

In this video, I’ll build a Social Media Factory workflow that turns a long-form article into posts for multiple social media platforms. I’ll start with a form where I enter a project title and the Google Docs article ID, then use the Google Docs node to retrieve the article content from Google Drive.

From there, I’ll add an AI Agent and build a structured prompt that asks the AI to repurpose the article for X, LinkedIn, and Facebook. I’ll define different rules for each platform, including tone, length, number of posts, emojis, hashtags, calls to action, and paragraph structure.

I’ll also ask the AI to create an image prompt for every post, with the correct aspect ratio for each platform. Once the posts are generated, I’ll save the finished social media content into a new Google Doc inside a Social Media Factory folder.

Finally, I’ll test the workflow with both a local model and a stronger OpenRouter model, comparing the quality of the output and showing how the image prompts can be used in ChatGPT to generate matching social media visuals.

This brief asks you to build a Resume and Cover Letter Assistant that takes a job listing and an existing resume, then uses AI to create tailored application documents for that specific role. The automation should analyze the job description, identify the key skills and requirements, compare them with the candidate’s real experience, and rewrite the resume without inventing qualifications or achievements. It should then create a customized resume and cover letter, and save the final documents somewhere useful, such as a Google Drive folder for each job application.

In this video, I’ll start building a Resume and Cover Letter Assistant workflow in n8n. The goal is to take one existing resume, compare it against multiple job descriptions, and automatically create a tailored resume and cover letter for each job application.

I’ll begin by setting up a Google Drive structure with a main Job Seeker Helper folder, a current resume, a Job Descriptions folder, and an Applications folder where the finished documents will be saved. Then I’ll create a form trigger that asks for the resume ID and the job description folder ID.

Next, I’ll use Google Docs to retrieve the resume and Google Drive to search the job description folder. For each job description found, the workflow will get the document content and pass it, along with the resume, into an AI Agent.

I’ll then build a detailed prompt that tells the AI to analyze the job listing, tailor the resume without inventing any experience, and write a matching cover letter. I’ll also add strict rules to prevent the AI from fabricating jobs, qualifications, achievements, or skills. Finally, I’ll ask the AI to return the tailored resume and cover letter as structured JSON, so we can separate and process them in the next part of the workflow.

In this video, I’ll continue building the Resume and Cover Letter Assistant workflow and finish the document generation process. I’ll start by taking the JSON output from the AI Agent and using an Edit Fields node to separate the tailored resume and covering letter into individual fields.

Then I’ll create a new application folder for each job description inside Google Drive, using the company name as the folder title. Inside each application folder, I’ll create two Google Docs: one for the customized resume and one for the covering letter.

After testing the first version, I’ll show why the raw AI output needs better formatting. To fix that, I’ll add two more AI Agents: one to format the resume professionally in Markdown, and another to format the covering letter with proper structure, spacing, date, greeting, body, and sign-off.

Finally, I’ll compare the results from a local model with a stronger OpenRouter model, showing how the quality and reliability of the final documents can vary depending on the AI model you use. By the end, the workflow will create a separate folder for each job application, complete with a tailored resume and covering letter ready for review.

This lecture includes all of the automation JSON files created in the Project Briefs section as Resource Downloads.

The lesson itself will show you how to import and use these.

I do recommend you build everything yourself as you follow along with the lessons in this course, and that means having a genuine attempt at all briefs yourself. But at the same time, I recognise the fact that my solutions will be different to yours, and you may like to explore my solutions as an extension to your studies.

This lecture explains what this section is about.

This video goes through Phase I, which includes the initial research up to a basic outline. All the details are written to Google Sheets as a kind of database.

This video shows how I improved the outline, and generated the articles.

This is phrase III of the automation, where we create a post and featured image, and then post everything to WordPress.

Some ideas on adding to this automation.

This lecture contains the automation download as a resource.

In this video, I’ll wrap up the course by showing how Comet Browser from Perplexity can help when you get stuck building n8n automations. I’ll explain how you can use it to generate prompts, troubleshoot errors, and get guidance on which nodes to use in a workflow.

I’ll also talk about an important limitation when using local AI models: if your prompt is too large or complex, smaller models may struggle to follow every instruction. I’ll show why it can be better to break a large task into smaller AI Agent steps, such as using separate agents for X, Facebook, and LinkedIn posts instead of asking one agent to handle everything.

Then I’ll give examples of how to ask Comet for help, including creating clean Markdown prompts, generating image prompt instructions, planning an n8n workflow, avoiding code-based solutions, and troubleshooting error messages. By the end, you’ll see how Comet can act as a useful support tool when you’re building, testing, and improving your own n8n automations.

Closing message...

This lecture includes a PDF showing how to install n8n on a VPS.