Connect Aircall to HubSpot with n8n: Auto-Log Calls and Trigger Follow-ups

The native Aircall HubSpot integration does one thing: it drops a call activity on the contact record. That is it. No outcome-based routing, no sequence enrollment, no custom property writes, no Slack ping when a rep tags a call “Interested.” For most RevOps setups I encounter, that means reps are manually logging follow-up tasks or those calls are vanishing into a black hole. This post is the workflow recipe that fixes it: an n8n automation that catches every Aircall event, writes enriched data to HubSpot, and fires the right follow-up sequence based on call outcome.

Why the Native Integration Falls Short

Aircall’s native HubSpot connector is a one-way pipe built for simplicity. It logs that a call happened. What it cannot do is read the call tag your rep applied, check whether the contact already has an open deal, and then conditionally enroll them in a “Warm Lead” sequence while simultaneously setting a custom property and pinging your Slack channel. That conditional branching is where middleware earns its place.

HubSpot’s Engagements API accepts call metadata including duration, disposition, recording URL, and notes. All writeable fields the native Aircall connector ignores entirely. Aircall’s webhook documentation exposes call.ended events with everything you need: caller number, duration in seconds, tags array, direction (inbound or outbound), and recording URL if your plan includes it.

My strong opinion here: don’t bother trying to patch the native connector with workarounds. Build the n8n layer once, build it right, and own the logic permanently.

The workflow I’m walking through runs on n8n’s self-hosted or cloud instance. If your team is on Make, the logic maps 1:1. You’ll use Aircall’s webhook module and HubSpot’s Update Contact plus Create Engagement modules instead of n8n nodes. I’ll flag the Make equivalents where they diverge.

The Workflow Architecture

Before jumping into steps, here is the shape of what we are building. Aircall fires a call.ended webhook to n8n. n8n looks up the caller’s phone number in HubSpot to find the matching contact. It writes a Call Engagement with duration, recording link, and notes. Then it branches on call tags: if the rep tagged the call “Interested,” enroll the contact in a follow-up sequence; if they tagged it “No Answer,” set a callback task for 24 hours out; anything else routes to a Slack alert for manual review.

The Gotcha That Will Break Your Enrollments

The most common failure point with this workflow is sequence enrollment. HubSpot’s Sequence Enrollment API is strict: the contact must have a valid email, the enrolling user must own a Sales Hub Professional or Enterprise seat, and the contact cannot already be actively enrolled in another sequence. If any of those conditions fail, the API returns a 4xx and n8n marks the execution as an error. Nothing tells your rep. The call logs correctly, the follow-up never fires, and you find out three weeks later when pipeline is short.

// n8n Code node: pre-enrollment guard
const contact = $input.item.json;
const email = contact.properties?.email;
const activeEnrollments =
  parseInt(
    contact.properties
      ?.hs_sequences_actively_enrolled_count ?? '0'
  );

if (!email) {
  return [{ json: {
    skip: true,
    reason: 'no_email',
    contactId: contact.id
  }}];
}

if (activeEnrollments > 0) {
  return [{ json: {
    skip: true,
    reason: 'already_enrolled',
    contactId: contact.id
  }}];
}

return [{ json: {
  skip: false,
  contactId: contact.id,
  email
}}];

Which Automation Platform Should Run This?

If you are deciding between n8n and Make for this specific workflow, the decision is less about features and more about how your team works day to day.

n8n vs Make for Aircall Automation

Choose n8n if

• You want self-hosted data handling so call recordings and phone numbers never touch a third-party cloud
• Your workflow needs complex branching logic or custom JavaScript transforms that click-and-drag builders make painful
• You are already running n8n for other RevOps automations and want one credential store

Free self-hosted or from $20/mo cloud Try n8n →

Choose Make if

• Your team is non-technical and needs to maintain the workflow without writing code
• You want a visual scenario builder where the Aircall to HubSpot logic is drag-and-drop with prebuilt modules
• You are already paying for Make and have operations headroom in your plan

Free tier, then from $9/mo Try Make →

Verifying the Workflow Works

Before you put this in production, run through four checks. Make a test call from an Aircall number that is already a HubSpot contact and verify the Call Engagement appears on the timeline within 30 seconds. Tag the call “Interested” and confirm the contact enrolls in your target sequence. Check the HubSpot sequence dashboard, not just the contact record. Make a call from a number with no HubSpot contact and confirm a Slack alert fires instead of a silent failure. Finally, check the “Last Call Duration” and “Last Call Outcome” custom properties on the contact record to confirm they populated correctly.

This workflow pairs naturally with the Apollo to HubSpot sync pattern I covered previously. If you are enriching inbound leads via Apollo before they hit a call with a rep, you can chain these two workflows so the contact arrives on the call with full context already written to their record.

If your team is evaluating where Aircall fits in a broader outbound stack, this breakdown of the best CRM setups for outbound teams covers how dialer integrations factor into the decision between HubSpot and its alternatives.

Putting It Together

This n8n workflow takes about three focused hours if you have your HubSpot credentials and Aircall API key ready. In my testing, the end-to-end call-to-sequence-enrollment fires in under 30 seconds from call end. The payoff is a call-to-follow-up loop that runs without rep input: every qualified call automatically logs, custom properties get written for reporting, and warm leads enter a sequence before your rep has closed their laptop.

The pre-enrollment guard is the piece most tutorials skip, and it is the piece that determines whether this workflow is reliable or just occasionally works. Build that guard first. Test it on real calls. Then layer in the Slack fallbacks so nothing drops silently. A workflow that fails loudly is fixable. One that fails quietly costs you pipeline.