n8n Workflow: Scheduled Competitor UX Audit

What This Builds

A scheduled workflow in n8n that captures screenshots of a competitor's key screens on a recurring basis, runs each screenshot through an AI vision model, and compiles the notes into a running audit document. Instead of remembering to manually check a competitor's product every month, the workflow checks for you and hands you a summary of what changed.

Prerequisites

• n8n account. The free self-hosted tier works for a low-volume workflow, or check n8n.io/pricing for the hosted cloud version
• A screenshot API or browser automation node (n8n has HTTP Request nodes that can call a screenshot service) to capture the competitor's public screens
• An OpenAI or Anthropic API key for the vision analysis step, billed per use, typically a few cents per screenshot
• A place to store the running audit, a Google Doc, Notion page, or shared spreadsheet
• Total ongoing cost: Free if self-hosted, or check n8n.io/pricing for the hosted cloud version, plus API usage, usually under $5 a month for a monthly audit of a handful of screens

The Concept

Picture a research assistant who visits a competitor's public website on the first of every month, takes photos of the pages you care about, and writes you a short note on anything that looks different from last time. That's what this workflow automates: a scheduled visit, a screenshot, an AI description of what's on the screen, and a running log you can scan in minutes instead of manually revisiting the competitor's site yourself.

Build It Step by Step

Part 1: Set Up the Schedule Trigger

1. In n8n, click the + button on the canvas and search for Schedule Trigger. Add it as the first node.
2. Configure it with a preset interval (for example, monthly) or a cron expression for a specific day and time.
3. Save the workflow. n8n only applies schedule changes after you save and publish, so an edited interval that isn't published will not run on the new schedule.

Part 2: Capture the Competitor Screens

1. Add an HTTP Request node connected to a screenshot API (several offer a free tier for a handful of monthly captures) pointed at the competitor's public URL, or use n8n's browser automation integration if you have one connected.
2. If you're tracking multiple screens (homepage, pricing page, onboarding flow), add a Split Out node feeding a list of URLs into the HTTP Request node so it captures each one in sequence.
3. Only capture public-facing pages a competitor makes available to any visitor. Do not attempt to access anything behind a login you don't have legitimate access to.

Part 3: Run the AI Vision Analysis

1. Add an OpenAI (or Anthropic) node configured for vision/image input, connected after the screenshot step.
2. Use a prompt like:

You're reviewing a screenshot of a competitor's product screen for a UX audit. Describe: (1) the primary layout pattern, (2) any UI pattern that differs from the last audit note provided below, (3) one thing worth testing on our own product if this represents a meaningful shift. Be specific about what's visible, don't guess at functionality you can't see in the image.

Previous audit note for this screen: {{previous note text, or "none, first audit"}}

1. Test this step against a captured screenshot to confirm the description is accurate and specific, not generic.

Part 4: Log the Findings

1. Add a Google Sheets, Notion, or Google Docs node to append the AI's notes to a running audit log, tagged with the date and screen name.
2. Structure the log with columns or sections for Date, Screen, AI Notes, and a manual "Reviewed" checkbox you fill in yourself after reading.

Part 5: Test and Publish

Run the workflow manually once (n8n lets you execute a workflow on demand for testing) before relying on the schedule. Confirm the screenshot captures the right screen, the AI note is coherent, and the log entry appears correctly formatted. Then publish the workflow so the schedule takes effect.

Real Example: Tracking a Direct Competitor's Onboarding Flow

Setup: A UX designer at a fintech startup set up a monthly n8n workflow tracking three screens on a direct competitor's public site: homepage, pricing page, and the first screen of their public signup flow.

Input: Nothing manual. the schedule fires on the first of the month and captures all three screens automatically.

Output: A running Google Doc log with one entry per screen per month. In one cycle, the AI note flagged that the competitor had added a progress indicator to their signup flow that hadn't been there the previous month, worth a quick look at whether the designer's own product's flow needed the same treatment.

Time saved: A manual competitor check like this, done properly with screenshots and notes, takes 30-45 minutes a month per competitor tracked. Automated, it takes the few minutes needed to read the log entry once it lands.

What to Do When It Breaks

• The workflow silently stops running and nobody notices for months → n8n workflows can fail quietly if a node's API key expires or a screenshot service changes its response format. Check the Executions log in n8n at least monthly, and if available on your plan, turn on an error-notification workflow that emails you when any execution fails.
• Screenshot comes back blank or shows a cookie banner → Add a short delay before capture, or configure the screenshot service to dismiss common cookie/consent overlays before capturing.
• AI description is vague or generic → Feed the previous audit note into the prompt (as shown in Part 3) so the AI has something concrete to compare against instead of describing the screen in isolation each time.
• Competitor changes their site structure and the URL 404s → Add a simple status-code check after the HTTP Request node so a failed capture logs an error entry instead of silently skipping that screen.

Variations

• Simpler version: Skip n8n entirely and set a monthly calendar reminder to manually screenshot the same screens and paste them into ChatGPT with the same analysis prompt. No automation, but no setup cost either.
• Extended version: Track three or four competitors in parallel branches of the same workflow, and add a summary step that compiles all of them into one monthly digest instead of separate logs.

What to Do Next

• This week: Build the workflow for one competitor and one or two screens, and run it manually to confirm it works end to end.
• This month: Let it run on schedule for a full cycle, then check the Executions log to confirm it actually fired.
• Advanced: Feed the monthly log into a chained summary prompt that produces a quarterly "what's changed in the market" brief for stakeholder reviews.

Advanced guide for UX designer professionals. These techniques use more sophisticated AI features that may require paid subscriptions.