Home Services Pricing About Us Apps & Tools Blog Get Started
AI Agents & Breeze, Configured Right

AI Agents & Breeze,
built into your HubSpot.

HubSpot's AI moved fast in 2026 — Breeze, Agent Hub, Agent Builder, outcome-based credits. We keep up with it so you don't have to, and we build the parts HubSpot doesn't ship out of the box.

In Plain English

What "Breeze" actually means
right now.

HubSpot renamed and reorganized this twice in the last year. Here's where it actually stands, no jargon.

Free, Built-In

Breeze Assistant

The AI helper inside HubSpot. Drafts emails, summarizes records, answers questions about your data. Free on every plan, but it only responds — it doesn't act on its own.

Autonomous Workers

Agent Hub

Formerly "Breeze Agents." These run on their own — the Customer Agent handles support tickets, the Prospecting Agent researches and reaches out to leads, the Data Agent cleans up records. Priced per completed task.

Build Your Own NEW

Agent Builder

Formerly "Breeze Studio." A no-code way to configure custom agents around your own workflows, knowledge base, and escalation rules — this is where most of the real value is for a specific business.

What We Build

Four ways we help you
put AI to work in HubSpot.

You don't need every agent turned on. You need the ones that fit how your team actually works, configured correctly, and watched so they don't burn credits on nothing.

Configuration

Set Up Core Agents Right

Customer, Prospecting, Data, and Company Research agents — configured against your actual knowledge base, pipelines, and escalation rules instead of the defaults.

Ask About Setup
Custom Build

Custom Agents in Agent Builder

Purpose-built agents for a specific job HubSpot's core agents don't cover — trained on your content, scoped to your process, with clear rules for when a human takes over.

Ask About Custom Agents
Our Specialty NEW

AI-Computed Properties

Custom HubSpot properties that fill themselves in — an AI prompt pulls data from other fields on the record and writes the answer back. Region, classification, a plain-English summary, whatever you need.

See How It Works
Hybrid Automation

N8N & Make, Where It Still Wins

Breeze doesn't replace everything. For automations that touch tools outside HubSpot, we still build in N8N or Make and connect it cleanly — no rip-and-replace just because HubSpot shipped a new feature.

Ask About Automation
Real Examples

Eight agents you can actually
turn on right now.

What each one does, why it's worth setting up, and which plan you need before we quote anything.

Pricing shown reflects HubSpot's published rates as of this writing and is set by HubSpot, not us — we'll confirm current numbers and what plan you're actually on before recommending anything.

Our Specialty

AI properties that
fill themselves in.

Instead of a person or a rigid formula deciding a property's value, a prompt looks at the other data already on the record and writes the answer — updated automatically as that data changes.

The Problem

Region & Segment Fields Are Always Manual or Wrong

Someone picks the region dropdown by hand, or a workflow uses rigid if/then logic that breaks the moment a new state or edge case shows up.

Our Solution

AI Region Property

Reads Address, State, Territory
Prompt Fills the Field
Updates on Record Change
Falls Back to Manual Review
Python + HubSpot APIai_region_property.py
def compute_ai_region(contact):
    context = {
        "state": contact.state,
        "zip": contact.zip,
        "territory_notes": contact.territory_notes,
    }
    prompt = build_region_prompt(context)
    region = call_ai(prompt)
    update_property(contact.id, "ai_region", region)

Region is one example. The same pattern works for lead classification, a plain-English "why this deal is stalled" summary, data normalization, next-best-action suggestions — anywhere a property should reflect several other fields instead of just one.

Automation Outside HubSpot

When it needs to leave HubSpot,
we build it in N8N or Make.

Breeze and Agent Builder are genuinely good at what happens inside HubSpot. The moment a workflow needs to touch several other tools or run logic HubSpot's automation builder wasn't built for, we still reach for N8N or Make — and connect it back to HubSpot cleanly.

Self-Hosted · Developer-Friendly

N8N

Open-source and self-hostable — the right call for high-volume workflows, custom code steps, or when a business wants to own its automation infrastructure.

No-Code · Fast to Ship

Make

A visual, no-code builder that's faster to stand up and easier to hand off to a non-technical team — the right call for straightforward multi-app automations.

CRM Record Sync SMS & Email Platforms Spreadsheets & Databases Custom & Third-Party APIs AI Model APIs Scheduling & Booking Tools

Not sure which one fits? Tell us what you're trying to automate and we'll tell you honestly whether it belongs in HubSpot, N8N, or Make.

Ask About Automation
The Right Order

Don't turn everything on
at once. Here's the sequence.

Most of the trouble we see isn't a bad agent — it's five agents switched on the same week with no data underneath them. This is the order we actually run for clients.

01

Breeze Intelligence (Enrichment) First

Every agent below performs better with clean, enriched data — and standard enrichment is free.

02

One Core Agent — Customer or Data

Pick whichever matches your actual bottleneck. Not both in the same week.

03

Prospecting or Company Research Agent

Once your first pilot has proven out the credit-monitoring habits your team needs.

04

Custom Agents (Agent Builder)

The highest-leverage tool here, and the most likely to go confidently wrong if rushed. Only after the review habit is in place.

05

AEO & Customer Health Agent (Beta)

Lower urgency for most businesses. Good additions once everything above is running smoothly, not a starting point.

Want the full playbook? Every setup step, best practice, and common mistake for all 8 agents — the same reference doc we use internally.

Get the Free Guide
Before You Scale

The credit-governance checklist
we run on every account.

Outcome-based pricing means a good month can look like a bad bill if nobody's watching. This is what we check before letting anything run unsupervised.

A maximum monthly credit limit is set at the account level
Someone specific owns reviewing AI usage and performance monthly
Every active agent is tied to a measurable outcome — hours saved, tickets deflected, leads booked
No agent was switched on and then abandoned after one odd result
Beta features (like Customer Health Agent) have a reminder set for when free access ends
Handoff and approval rules are reviewed, not left on default
Get Started

Ready to see what AI
actually fits your portal?

Tell us what you're trying to automate. We'll tell you honestly whether it's a Breeze agent, a custom build, or still N8N under the hood.