HubSpot AI agents, configured and extended for your team

What AI agents HubSpot offers

HubSpot groups its AI features under Breeze. The agents include a Customer Agent that answers support questions from your knowledge base across chat and email, a Prospecting Agent that researches contacts and drafts outreach, and agents for content, social posts and knowledge base articles. HubSpot has also introduced a Breeze Studio area for configuring agents and a marketplace for additional ones, and the lineup changes often, so check HubSpot's current product pages for your portal.

Alongside agents, HubSpot offers AI assistants inside the editor and CRM, predictive lead scoring on some tiers, and classic workflow automation. Access to agents typically requires Professional or Enterprise subscriptions, and usage is metered in HubSpot credits. The practical question is not which agents exist, but which ones match a problem you already have and have the data they need.

How to set up an AI agent in HubSpot and what to automate first

Setting up a HubSpot AI agent follows the same pattern for each one. Confirm your tier and credit allowance, then clean the data the agent will read, such as contact properties, lifecycle stages and company records. For a Customer Agent, publish accurate knowledge base articles and define when to hand off to a human. For a Prospecting Agent, define target segments, tone and which sequences it may enrol people into. Test in a limited channel before opening it widely.

For HubSpot automation workflows, start with the unglamorous ones: lead assignment and rotation, lifecycle stage updates, deal stage hygiene, task creation after meetings and renewal reminders. These HubSpot workflow automation examples are rule-based and do not need AI. Add agents where judgement or language is involved, such as triaging support tickets, summarising calls into the record or researching new inbound accounts.

Building a custom AI agent for HubSpot and what it costs

HubSpot's own agents work well inside HubSpot. A custom agent makes sense when the job spans systems, for example pulling billing status from Stripe or NetSuite before a renewal, reading product usage from your database, or applying a routing policy too complex for workflows. Custom agents connect through the HubSpot API or a HubSpot app, write back to properties, notes and tasks, and can request a person's approval in Slack before any email or deal change.

On cost, HubSpot's agents are included with certain tiers but consume credits, and heavy usage can require buying more. A custom HubSpot agent is a separate build cost plus model usage, which is usually modest per task. Build scope is driven by the number of connected systems and approval steps. Velum scopes the work, recommends native Breeze agents where they are enough, and quotes a fixed price for anything custom.

How it works

1. 1

   Audit the portal

   We review your tier, credits, properties, workflows and data quality to see which native agents are usable today.

2. 2

   Clean the data agents depend on

   We fix lifecycle stages, required properties, duplicates and knowledge base gaps so agents read accurate information.

3. 3

   Configure native Breeze agents

   We set up the HubSpot agents that fit your needs, with defined tone, scope and handoff rules.

4. 4

   Build custom agents for the gaps

   Where a job needs other systems or complex logic, we build a custom agent that works through the HubSpot API and logs everything to the record.

5. 5

   Pilot with approvals

   We launch to one team or channel first, with a person approving outbound emails, deal changes and deletions until results are steady.

Before and after

Typical ranges from comparable deployments. Your baseline is measured before anything is built.

Tools it works with

• HubSpot
• HubSpot Breeze
• Salesforce
• Stripe
• NetSuite
• Gmail
• Microsoft 365
• Slack
• Zapier
• Claude