n8n HubSpot CRM: Auto Create Contacts and Deals

Every small business I talk to has the same leak somewhere in their pipeline. A form gets filled out, a lead replies to an email, someone books a call, and none of it lands in HubSpot until a human remembers to type it in. An n8n HubSpot CRM workflow closes that gap. It watches for the trigger you care about and creates the contact and the deal the second the event happens, not whenever someone gets around to it.

What an n8n HubSpot CRM workflow actually does

Strip away the buzzwords and it is simple. n8n listens for something (a new form submission, a Calendly booking, a Stripe payment, an inbound email), pulls the relevant fields, and calls the HubSpot API to create or update a contact, then attaches a deal to it. You are not replacing HubSpot. You are just removing the part where a person sits between the event and the record.

Building the contact and deal flow

The core pattern I use looks like this in n8n:

• A trigger node for the source (webhook, form tool, or a scheduled poll)
• A HubSpot node set to search for the contact by email first
• A branch that updates the existing contact or creates a new one
• A second HubSpot node that creates a deal and associates it with that contact
• A Slack or email notification so your sales person actually knows something new landed

That search step matters more than people think. Skip it and you end up with three duplicate contacts for the same person because they filled out two forms and replied to an email. Always check for an existing contact before you create one. It is the single biggest quality difference between a workflow that looks good in a demo and one that survives six months of real traffic.

Where teams get this wrong

The mistake I see most is trying to map every HubSpot field on day one. Start with the fields that actually drive your sales process: name, email, source, and deal value. You can always add lifecycle stage or lead score later once the basic n8n HubSpot CRM flow is proven stable. I also recommend logging every run to a simple sheet or database table for the first few weeks. When something breaks, and something always breaks eventually, usually a rate limit or a renamed field, you want a paper trail instead of guessing.

The Esipick angle

We built almost this exact workflow for our own pipeline before we ever sold it to a client. Leads used to come in through three different channels and sit unrecorded for a day or two while I meant to enter them manually. Once we wired n8n to HubSpot with the search before create logic above, our response time to new leads dropped noticeably, not because the automation is clever, but because nothing sits in a queue anymore. The lesson that stuck with me: automation does not need to be sophisticated to matter. It just needs to run the moment the event happens instead of whenever a human remembers.

If your team is still copying leads from one tool into another by hand, that is hours a week you are not getting back, and it is one of the fastest workflows to fix once you actually see it.