Agent Hub

Freemium

Agent Hub is an AI agent management platform for building, running, and watching AI agents across marketing, sales, and customer service. It plugs straight into your HubSpot CRM and uses that data as its brain.

You get six pre-built agents. Content Agent writes blog posts and social copy. Campaign Agent plans multi-channel campaigns from one goal. Prospecting Agent watches for buying signals and reaches out. Customer Agent answers support tickets 24/7 using contact history. Data Agent answers custom questions about your contacts. Nurture Agent personalizes email follow-up. You can also build custom agents in a no-code builder.

Everything runs from one dashboard where you see agent status, activity, and what they produced. All of it feeds off your CRM data, for better or worse.

Open it if you are already on HubSpot Professional or Enterprise with clean CRM data and one person who owns the business rules. Skip it if you are on Starter, your data is messy, or you need agents that talk to tools outside HubSpot.

The fit is narrow. These are the cases where it actually works.

• Teams already on Professional or Enterprise tiers with clean CRM data: The agents only know what your CRM knows. Dirty data means dumb agents.
• Organizations that can name one specific owner for the Context tab: This tab holds your company voice, ICP definitions, and knowledge. If nobody owns it, every agent inherits the mess.
• Use cases where volume is predictable enough to model credit costs: Credits burn fast. You need to forecast before you deploy.
• Teams drowning in manual lead qualification and routing: Prospecting Agent can take that work off your plate if your rules are clear.
• Customer service operations with high volume of repetitive tickets: Customer Agent handles tier-one support if you have documented answers in HubSpot.
• Sales teams needing systematic deal reviews with context beyond simple rules: Deal Review Agent flags stale deals and missing buyer commitments.
• Marketing teams ready to shift from execution to agent supervision: You stop writing drafts and start checking what the agent produced.
• Users willing to configure and test before production use: Simulated testing costs nothing. Skipping it costs you mistakes at scale.

These are the parts worth the trouble after you get past setup.

• Single console for all agent activity: You see what is running, who turned it on, and where credits went. No more mystery agents.
• Context tab forces central ownership: One place for voice, rules, and knowledge means agents stay consistent instead of drifting.
• Simulated testing before live runs: You can iterate on custom agents without burning credits. The estimate shows before you commit.
• Structured outputs with evidence: Deal Review Agent does not just say "at risk." It cites the stale data and names the missing buyer.
• Clear credit pricing lets you model costs: You know the burn rate before you scale, not after the bill arrives.

The path from zero to running an agent.

1. Go to the Agent Hub console in your HubSpot account.
2. Define your Context tab: company voice, ICP, and knowledge vault.
3. Pick a pre-built agent from the marketplace or start a custom build.
4. Test in simulation mode until the output looks right.
5. Activate live runs and watch the Agent Inbox for status and credit burn.

Skip it when these limits match your situation.

• Starter tier lockout: Professional or Enterprise minimum creates a hard floor at $450+ per month.
• Credit costs escalate unpredictably: 50 credits per resolved conversation sounds small until volume spikes.
• Sandbox accounts cannot test credit actions: You cannot fully evaluate before you pay.
• Documentation is already stale: Renamed from Breeze in July 2026. Old guides send you to dead links.
• Cloning makes agent sprawl easy: Governance is manual. Nobody follows the approval workflow.
• Knowledge trapped in HubSpot only: Agents cannot read your Notion, Slack, or external help docs.