DEV Community

Sasidhar Sunkesula
Sasidhar Sunkesula

Posted on

I built an n8n workflow that watches my competitors so I don't have to

#automationn8ncompetitors

Most agencies and SaaS teams check their competitors' pricing and feature pages manually. Once a week. If they remember.

It's tedious, easy to miss, and by the time you notice a competitor dropped their prices or launched a new feature, you've already lost ground.

I built an n8n workflow that watches competitor pages automatically and only alerts me when something actually changes.

What it does

Every day at 7 AM, the workflow:

  1. Fetches each tracked competitor's page (pricing, features, homepage) via HTTP
  2. Extracts the text content and computes a content hash
  3. Diffs against the last snapshot — if nothing changed, it skips silently (no noise)
  4. AI summarizes what changed and the business impact (only on real changes)
  5. Alerts me on Slack with the change summary + recommended action
  6. Saves the new snapshot for next comparison

The key insight: hash-based dedup

The magic is in the diffing. Most "competitor monitoring" tools alert you on every fetch, which means alert fatigue. You stop paying attention.

This workflow computes a content hash (djb2) and only alerts when the hash changes. Identical content = silent. Real change = alert.

No noise. Only signal.

The architecture 

Schedule (daily 7AM)
  → Competitor Targets (Sheets or demo)
  → Validate Target (URL, page type)
  → [Invalid?] → Log to "Competitor Errors"
  → HTTP: Fetch Page
  → Extract + Hash + Diff (djb2 hash)
  → [No change?] → Stop silently
  → Gemini: Summarize Change (what changed + business impact)
  → Parse Summary (with fallback)
  → Slack: Change Alert
  → Sheets: Save Snapshot
Enter fullscreen mode Exit fullscreen mode

Production features

  • 3× retry on every external node
  • Hash-based idempotency — won't re-alert on identical content
  • Confidence gate — low-confidence summaries flagged "verify before acting"
  • Error branches — invalid targets and parse failures route to error logs
  • Snapshot history — every change logged with hash for audit trail

The result

I get a Slack message only when something actually changes. No daily noise. No manual checking. Just signal.

Setup: $400–700. Monthly: $150–250.

This is competitive intelligence on autopilot.

I build production n8n automations with AI integration. 13 workflows across 8 domains, all tested end-to-end. GitHub | LinkedIn

Top comments (0)