New Pipedrive organizations get an AI note and a Slack post

New Pipedrive organizations get an AI note and a Slack post

An n8n workflow that reacts when someone creates an organization in Pipedrive. It reads the organization's website
through ScrapingBee, has an AI model write a short company summary, saves it as a note on the organization and posts it in
a Slack channel.

What is inside

• workflow.json: the workflow as n8n exports it, 7 nodes and 1 sticky note. It holds no credential.
• README.md: these notes.

What we changed from the original copy

• The note was written by gpt-4o. OpenAI's deprecations page, read on 12 November 2026, shuts down the dated gpt-4o
  snapshots on 14 January 2027 and names gpt-6-luna as the replacement. The node now names gpt-6-luna. It only sends
  a prompt and has no tool connected, so that model fits. OpenAI's model page, read on 13 November 2026, lists a context
  window of 1,050,000 tokens, which helps with long pages of HTML.
• The trigger was a version 1 node. Pipedrive's API reference says a webhook made without a version is version 2, the
  default from 17 March 2025. The version 1 node sends no version, so the events arrive in the version 2 shape while the
  expressions expected the old one. The trigger is now version 1.1 of the node, set to the
  create action and the organization entity.
• The expressions now read the version 2 body. The website is in data.custom_fields, under the key of your field, as an
  object with a value. The organization id is meta.entity_id, which Pipedrive describes as the id of the object, and
  the name is data.name. The old copy read current and meta.id.
• The key of the website field was a placeholder, <random_api_id_custom_website_field>. It is now
  <YOUR_WEBSITE_FIELD_KEY>, and the ScrapingBee key stays a placeholder, <YOUR_SCRAPINGBEE_API_KEY>.
• The nodes had long names that carried a model name and a spaced hyphen. They are now New Pipedrive Organization,
  Fetch Website Page, Summarize Page with AI, Add Note to Organization, Markdown to Slack Format and
  Notify Slack. The Slack text had the typo "Organizaton", now fixed, and the cached channel name is gone.
• The sticky note is rewritten to match these changes, without links.

Before you start

• An n8n instance reachable from the internet over https, on n8n Cloud or your own server. Pipedrive calls the webhook
  that the trigger creates.
• A Pipedrive account with a custom text field for the website on organizations, and a Pipedrive credential in n8n. Pipedrive
  checks each webhook event against the permissions of the user it runs as, so that user must be able to see the
  organizations.
• A ScrapingBee API key. ScrapingBee's page offers 1,000 free credits to start.
• An OpenAI API key.
• A Slack app with an OAuth2 credential in n8n. The token needs the chat:write scope. The app must be in the channel, or
  also hold chat:write.public.

Set it up

1. In n8n, create a workflow, open its menu, choose Import from File and pick workflow.json.
2. Add a Pipedrive credential to New Pipedrive Organization and to Add Note to Organization.
3. Look up the key of your website field in Pipedrive's organization fields API. In Fetch Website Page, replace
   <YOUR_WEBSITE_FIELD_KEY> with it and <YOUR_SCRAPINGBEE_API_KEY> with your key.
4. Add an OpenAI credential to Summarize Page with AI.
5. Add a Slack credential to Notify Slack and choose the channel.
6. Save the workflow and switch it on. Create an organization with a website in Pipedrive, and look for the note and the
   Slack post.

How a run goes

1. Pipedrive calls n8n when an organization is created, and New Pipedrive Organization passes the body on.
2. Fetch Website Page asks ScrapingBee for the page at the website address. It runs with render_js=false, so the page's
   scripts are not run and the call costs 1 credit.
3. Summarize Page with AI sends the HTML to the model. The instructions ask for a company description, the target market,
   selling points and possible competitors, written as HTML.
4. Add Note to Organization writes that text into Pipedrive as a note attached to the organization.
5. HTML To Markdown and Markdown to Slack Format turn the note into Slack's formatting. Level 1 and 2 headings become
   bold lines, star bullets become arrows, and the first table becomes lines of name: value.
6. Notify Slack posts the organization's name and the note.

Change the look

• What the note says: the system message in Summarize Page with AI.
• Whether the page's scripts run: render_js in Fetch Website Page. ScrapingBee's own default is on, at 5 credits.
• The model: Summarize Page with AI.
• The layout of the Slack post: the text of Notify Slack and the code in Markdown to Slack Format.
• The event that starts it: New Pipedrive Organization.

Good to know

• The facts below come from Pipedrive, ScrapingBee, OpenAI, Slack and n8n pages and source read on 12 and 13 November 2026.
  Nothing was run: no webhook was delivered, no page was fetched, no model call was made, and no note or Slack post was
  written.
• Scraping: ScrapingBee reads the page for you, and its terms bar use that breaches a scraped site's own terms. Follow the
  rules on web scraping in your country and the terms of each site, and remember that the address is whatever someone
  typed into the field.
• Only the create event starts the run, so the website must be filled in when the organization is created. An
  organization with no website leaves the fetch without an address, and the run ends in an error there.
• With render_js=false a page that builds its content in the browser may return little. Set it to true for those, at a
  higher credit cost.
• The whole HTML goes to the model. Long pages use many tokens, which is cheap on gpt-6-luna, USD 0.1 per million input
  tokens and USD 0.5 per million output tokens on OpenAI's page.
• The page text reaches the model as it is, so a page can hold words that try to steer the answer. The result goes straight
  into your CRM and Slack, so read the first notes before you trust them.
• The ScrapingBee key sits in plain text in the node, visible to anyone who can open the workflow. Use a credential instead
  if many people can.
• Your data: the website address goes to ScrapingBee, the page's HTML goes to OpenAI, and the note goes to Pipedrive and
  Slack.
• Credentials: the JSON holds none.