HubSpot Breeze Workflows: Step-by-Step Processes That Save Time

It helps to know what each component does before building workflows:

• Breeze Copilot: a chat assistant available across HubSpot that can summarise records, draft emails and answer questions about CRM data.
• Breeze Agents: specialised agents, such as the customer agent for support conversations, a prospecting agent for outreach research, and content and social agents.
• Breeze Intelligence: company and contact enrichment, buyer intent signals and form shortening using HubSpot’s data.
• AI features inside tools: email drafting, blog and social content suggestions, call summaries and workflow building assistance.

Many of these features use HubSpot credits, and some agents are tied to specific hubs or tiers. Admins can manage AI access and data settings in the account’s AI settings area.

Reps often spend the first part of their day reading company websites and old email threads. Breeze Copilot can do the first pass.

• Open the company or deal record and ask Copilot for a summary of recent activity, open deals and last conversations.
• Ask for a short company brief using enrichment data from Breeze Intelligence.
• Request three discovery questions based on the industry and the deal stage.
• Save the summary as a note on the record so the team can see it.

The rep still reviews and adjusts, but the research step shrinks. The quality depends on how well the CRM is maintained: if emails are not logged, the summary has gaps.

Marketing teams can use the content tools to produce first drafts of blog posts, landing pages, social posts and marketing emails. The key is giving Breeze context about the brand. HubSpot lets you set brand voice and business details that AI features use, so fill those in first.

A simple process works well. Create an outline with the content agent or Copilot, edit it, then generate section drafts. Run every draft through the same human editing and fact-checking you use for any copy. Repurpose the final article into social posts and an email using the built-in tools, so variations stay consistent with the approved text.

The customer agent can answer questions in chat and other channels using your knowledge base, website pages and documents you connect. It hands conversations to a human when it cannot answer or when the customer asks for a person.

Setting it up step by step

• Audit the knowledge base: remove outdated articles and fill gaps for common questions.
• Connect the sources the agent may use and exclude anything internal.
• Define handoff rules, such as billing disputes, complaints and any request for a human.
• Write a clear greeting that says the customer is talking to an AI assistant.
• Test with real past questions before going live, and review transcripts weekly.

Track resolution rate, handoff rate and customer satisfaction. If handoffs are high for a topic, the knowledge base is usually missing an article.

Incomplete records weaken every other workflow. Breeze Intelligence can enrich company records with details such as industry, size and location, and can shorten forms by filling known fields automatically. Combine that with HubSpot’s data quality tools, which flag duplicates and formatting issues.

Set a routine: a weekly enrichment run for new companies, a monthly duplicate review and workflow rules that set lifecycle stage and owner automatically. Check enrichment results on a sample before trusting them across the database, especially for smaller companies that may have less public data.

HubSpot’s workflow tool includes AI assistance that lets you describe an automation in plain language and get a draft workflow. For example: “When a contact submits the demo form and their company has more than 50 employees, assign to the enterprise team and send a confirmation email.” Review every trigger, branch and action before switching it on, and test with a sample contact. AI-drafted workflows are a starting point, not a finished build.

Breeze works on data inside HubSpot and the sources you connect. It does not know what happens in tools that are not integrated. Credit usage can grow quickly if many people run agents frequently, so monitor consumption in the account settings. AI outputs need review, especially customer-facing answers and content. Finally, permissions matter: make sure users only access AI features and records appropriate to their role.