n8n Marketing Workflows: 5 Blueprints to Copy (2026)

What n8n Costs: Verified 6 September 2026

• Starter — €20/month billed annually (free trial, no credit card): 2,500 workflow executions per month, 5 concurrent executions, 1 shared project, 2,300 AI credits per month, forum support. Runs all five blueprints at modest volume.
• Pro — €50/month billed annually: 10,000 executions per month, 20 concurrent executions, 3 shared projects, up to 13,700 AI credits, plus workflow history, execution search, global variables, and admin roles. Our pick for a team of three to five.
• Business — €667/month billed annually, aimed at companies under 100 employees: 40,000 executions, SSO/SAML/LDAP, Git-based version control, separate environments, 30 days of insights, and an AI Assistant in preview. The same page notes a Start-up plan with 50% off Business for teams under 20 employees.
• Enterprise — custom pricing through sales, hosted by n8n or self-hosted, with unlimited shared projects, 200+ concurrent executions, log streaming, extended data retention, and SLA-backed support.
• AI model options: the OpenAI node takes a key from platform.openai.com, and Anthropic models plug in through n8n’s Anthropic Chat Model node. Claude Pro is $20/month billed monthly or $17/month billed annually (anthropic.com/pricing, fetched 6 September 2026). More in our n8n and Claude profiles.

Blueprint 1: The Daily AI Brief (Feeds → Summary → Slack)

• Node 1 — Schedule Trigger: weekdays at 08:30 (cron 30 8 * * 1-5) in your team’s timezone. Test downstream with a Manual Trigger first; a schedule fires whether or not the workflow is correct.
• Node 2 — RSS Read: one node per feed — competitor blogs, industry newsletters, a Google News RSS search for your niche — each limited to 5 items.
• Node 3 — Item Lists: merge the feeds, sort by publish date, and cap the batch at 15 items so the AI step has a bounded input and predictable cost.
• Node 4 — AI step (OpenAI node or equivalent): system prompt “You are a marketing editor for {brand}, read by {audience}. From the items, choose the 3 most relevant. For each: headline, source, one line on why it matters to us, and one suggested take. Never invent details.” Set the model and temperature in the node; paste the prompt as-is, then tighten the output format from what comes back.
• Node 5 — Slack: post the brief to #marketing-brief as a formatted message, or send it with Gmail. End the run by appending the item titles to a Google Sheet so the next run can filter them out — otherwise day two repeats day one.
• Watch-outs: feeds break silently, so attach an Error Trigger that messages you when RSS Read fails; and the AI summary is for humans, not for republishing — run one [VERIFY]-style pass before anyone quotes it.

Blueprint 2: Lead Capture → AI Enrichment → Routing

• Node 1 — Trigger: the Typeform trigger on submission, or a Webhook node if your form tool can POST JSON. Capture name, work email, company, and message.
• Node 2 — HTTP Request to Tavily: POST to api.tavily.com/search with a JSON body containing your API key, the query “{company} about products” and max_results 5. Tavily’s free tier is enough to test; the call returns a JSON summary of the top results (Tavily profile).
• Node 3 — AI step: prompt “Classify this lead: company {company}, message {message}. Output JSON with segment (SMB, mid-market, enterprise), likely use case, buying signal 1-5, and a 3-line personalized opener a salesperson could send. Base every line only on the data provided.”
• Node 4 — If node: route on the buying-signal score. Strong leads continue to the CRM branch; the rest drop to a nurture sheet so the sales channel stays quiet.
• Node 5 — HubSpot node: create or update the contact and attach the enrichment summary as a note. Then a Slack node posts “New lead: {company} — {segment}, signal {score}” plus the opener to #sales-leads.
• Watch-outs: every enrichment call costs an API call or AI credit, so cap daily runs; and never auto-send the AI opener — routing to Slack exists precisely so a human decides in context.

Blueprint 3: Competitor Content Watch (Sitemap → Diff → Digest)

• Node 1 — Schedule Trigger: weekly, Monday at 09:00.
• Node 2 — HTTP Request: fetch each competitor sitemap (the XML at /sitemap.xml or /sitemap_index.xml). Parse the response with the Extract From File node set to XML, then flatten the URL list with Item Lists.
• Node 3 — Google Sheets: read the Seen URLs column from your tracker sheet and Filter out anything already logged, leaving only new and modified URLs.
• Node 4 — Sort + Limit: keep the ten most recent new URLs so a competitor publishing 80 pages in one week cannot flood the AI budget.
• Node 5 — AI step: for each new URL, fetch the page (HTTP Request plus Extract From File for HTML) and prompt “Summarize this page in one paragraph: topic, target keyword, angle, and the offer. Then say what we should do in response, or say nothing.”
• Node 6 — Output: append the new URLs to the tracker sheet and post the digest to #content, or email it with Gmail. Watch-outs: fetch only pages you are legally allowed to access and respect robots.txt and the site’s terms — the goal is competitive awareness, not scraping content to republish.

Blueprints 4 and 5: Draft Queues and the Weekly GA4 Digest

• Blueprint 4, Node 1 — Google Sheets trigger: runs on new rows in the Ideas tab. Expected columns: topic, target keyword, links, notes.
• Blueprint 4, Node 2 — Loop Over Items: processes each new idea separately, so five ideas cost five small AI calls instead of one giant prompt.
• Blueprint 4, Node 3 — AI step: prompt “Using the brand voice document at {link}, write for this topic: a LinkedIn post under 250 words, a 5-post X thread, a newsletter intro, and a carousel outline. Wrap any factual claim in [VERIFY].” Append the outputs to the Drafts tab with the status To review.
• Blueprint 4, Node 4 — Slack: notify #content that drafts are ready. The reviewer edits in the sheet and flips the status to Approved.
• Blueprint 4, Node 5 — publisher workflow on a Schedule Trigger every 15 minutes: read rows with status Approved, POST each to Buffer’s API (api.bufferapp.com) with your Buffer access token, mark the row Published, and raise an error if the API rejects the post.
• Blueprint 5, Node 1 — Google Analytics node: connect GA4 with OAuth2 and run a report for the last 7 days and the previous 7 days — sessions, users, top pages, conversions.
• Blueprint 5, Node 2 — AI step: prompt “Compare these two periods in plain language for a busy CMO. Three findings, one risk, one recommendation. No jargon.”
• Blueprint 5, Node 3 — Gmail: send the summary to the team every Friday and append the raw numbers to the reporting sheet for the quarterly review.
• Watch-outs for both: approve-by-default kills the review step, so only Approved rows ever publish; and GA4 sampling can make small-property numbers unstable, so treat the digest as a direction signal, not an audit.

FAQ: n8n Marketing Workflows