HubSpot Breeze Data Agent Guide: Setup & Workflows 2026

The HubSpot Breeze Data Agent is an AI-powered workflow tool that enriches, analyzes, and standardizes CRM data without manual effort. If your team spends hours researching companies, filling in missing contact fields, or cleaning up inconsistent records, the HubSpot Breeze Data Agent automates those tasks at scale using three dedicated workflow actions. The official Breeze Data Agent documentation covers each action’s full parameter set.

This guide walks through everything you need to set up and use the Breeze Data Agent effectively - from initial configuration through advanced chaining techniques that combine multiple AI actions in a single workflow. For the broader Breeze suite, start with our HubSpot Breeze AI getting started guide.

What Is the HubSpot Breeze Data Agent

The HubSpot Breeze Data Agent is one of four AI agents in the Breeze suite (alongside the Prospecting Agent, Customer Agent, and Content Agent). Unlike the other agents that focus on outbound sales, customer support, or content creation, the Data Agent focuses specifically on CRM data quality and enrichment within your workflows.

The Data Agent operates through three distinct workflow actions that you can add to any HubSpot workflow:

Custom Prompt Action

This action lets you write a natural language prompt (up to 3,000 characters) that the AI processes against your CRM record data. You might ask it to summarize a contact’s engagement history, categorize a company by industry vertical, or generate a personalized talking point for a sales rep. The AI uses an offline LLM - meaning it does not have internet access during execution - so it works exclusively with data already in your CRM.

Research Action

The Research action instructs the Data Agent to compile information about a contact or company using available CRM data and associated records. This is useful for building account briefs, identifying cross-sell opportunities, or flagging accounts that match your ideal customer profile based on existing data points.

Fill Smart Property Action

Smart Properties are AI-powered custom properties that automatically populate based on rules you define. The Fill Smart Property action triggers the AI to evaluate a record and fill in the designated smart property value. This is particularly powerful for lead scoring enrichment, industry classification, and data standardization across large contact databases.

Breeze Intelligence vs. Data Agent - An Important Distinction

A common point of confusion is the difference between Breeze Intelligence and the Breeze Data Agent. Breeze Intelligence is a separate product that enriches your CRM with external third-party data - pulling in company revenue figures, employee counts, technographic data, and buyer intent signals from outside sources. Our HubSpot Breeze Intelligence buyer intent guide covers Intelligence in depth. The Data Agent, by contrast, works with data that already exists inside your HubSpot CRM. It analyzes, categorizes, summarizes, and transforms your internal records using AI. Think of Breeze Intelligence as bringing new data in, and the Data Agent as making your existing data more useful.

When to Use the Breeze Data Agent

The Data Agent is not a replacement for manual data entry on individual records. It is designed for systematic, repeatable data operations that run across hundreds or thousands of records through workflows. For broader CRM data quality strategies, our best AI CRM tools 2026 roundup covers the wider tooling landscape. Here are the scenarios where the Data Agent delivers the most value.

Lead Scoring Enrichment

Use the Custom Prompt action to analyze a contact’s behavior history and generate a qualitative score or summary. For example, prompt the AI to evaluate whether a lead matches your ideal buyer persona based on their form submissions, page views, and email engagement. The HubSpot lead scoring documentation covers how AI scores integrate with native scoring properties. Store the result in a custom property that your sales team can filter on.

Company Research at Scale

When new companies enter your CRM through form submissions or imports, the Research action can automatically compile an account brief. The Data Agent pulls together associated contacts, deal history, support tickets, and engagement data into a structured summary - similar in spirit to how Gong and Outreach generate account intelligence. Sales reps get a complete picture without manually clicking through multiple record tabs.

Data Cleanup and Standardization

Inconsistent data is one of the biggest CRM problems - Gartner research consistently identifies data quality as a top barrier to effective CRM use. Job titles might appear as “VP Sales”, “Vice President of Sales”, “VP, Sales”, and “Sales VP” - all referring to the same role. Use the Custom Prompt action in a workflow to standardize these values. Prompt the AI to categorize each title into a normalized format and store the result in a standardized property.

Smart Property Population

When you need to classify records at scale - such as tagging every company with an industry vertical, identifying decision-maker contacts, or scoring account health - the Fill Smart Property action handles this automatically. The HubSpot properties documentation covers Smart Property setup. Set up the smart property with your classification rules, then trigger the workflow on new or updated records.

Workflow-Based Data Enrichment

Combine the Data Agent with enrollment triggers to create always-on enrichment pipelines. For example, when a contact submits a demo request form, a workflow could run the Research action to compile their account brief, use a Custom Prompt to assess fit against your ICP, and fill a Smart Property with a priority tier - all before the lead reaches a sales rep.

Setting Up Data Agent Workflows

Setting up your first Data Agent workflow requires a Professional or Enterprise HubSpot subscription. Navigate to Automation, then Workflows, and create a new workflow. The Data Agent actions appear under the AI section of the action menu. The HubSpot workflows knowledge base covers the underlying workflow builder mechanics in detail.

Step 1: Create a New Workflow

Navigate to Automation in the top navigation bar and select Workflows. Click “Create workflow” and choose your object type. The Data Agent works with Contact, Company, Deal, and Ticket objects. Select “Blank workflow” to start from scratch, then set your enrollment trigger. A common starting point is “Contact property is known” for the email field, which catches new contacts as they enter your CRM.

Step 2: Add a Custom Prompt Action

Click the plus icon to add an action and scroll to the AI section. Select “Custom Prompt” from the Breeze Data Agent options. In the prompt field, write clear instructions for what you want the AI to do with the record. Keep these guidelines in mind:

• The prompt limit is 3,000 characters - be specific but concise
• Reference CRM properties by name so the AI knows which data to use
• Specify the exact output format you want (for example, “Return one of: High Priority, Medium Priority, or Low Priority”)
• Remember that the AI runs offline with no internet access - it can only work with CRM data

Step 3: Add a Research Action

Add another action and select “Research” from the Data Agent options. Configure which aspects of the record you want the AI to research. You can specify focus areas such as engagement history, deal pipeline position, or associated company data. The Research action output is stored as a text property that you designate during configuration.

Step 4: Add a Fill Smart Property Action

Before using this action, you need to create a Smart Property first (covered in the next section). Once your Smart Property exists, add the Fill Smart Property action to your workflow. Select the target Smart Property from the dropdown. The AI evaluates the record against the property’s rules and populates the value automatically.

Step 5: Configure Output Properties

For each Data Agent action, specify where the result should be stored. Create dedicated custom properties for AI outputs - do not overwrite standard HubSpot properties. Use a naming convention like “AI - Lead Priority” or “AI - Account Brief” so your team can easily identify AI-generated values.

Step 6: Set Enrollment and Re-enrollment

Decide whether records should pass through the workflow once or re-enroll when data changes. For data cleanup workflows, a single enrollment is usually sufficient. For ongoing enrichment (like updating account briefs), enable re-enrollment with a time delay to avoid unnecessary credit consumption.

Each Data Agent action consumes 10 Breeze Credits per operation - see the official Breeze credits reference for current rates. Plan your workflows with credit usage in mind, especially when running against large contact databases.

Working with Smart Properties

Smart Properties are a feature unique to the Breeze Data Agent that lets you create custom properties powered by AI classification rules. Unlike standard HubSpot properties where values are entered manually or through form submissions, Smart Properties are populated by the AI based on criteria you define.

Creating a Smart Property

Navigate to Settings, then Properties. Click “Create property” and select the “Smart Property” type. Choose the object type (Contact, Company, Deal, or Ticket) and configure the following:

• Property name: Use a descriptive name like “AI Industry Classification” or “AI Lead Tier”
• Field type: Select from single-line text, dropdown, or number based on your use case
• AI instructions: Write the classification rules the AI should follow. For example: “Classify this company into one of the following industries based on their website domain, associated contacts’ job titles, and deal products: SaaS, E-commerce, Healthcare, Financial Services, Education, or Other”
• Options (for dropdown type): Define the allowed values the AI can assign

Configuring AI Fill Schedules

Smart Properties can be set to fill automatically on a schedule or triggered through workflows. The schedule option runs the AI classification across all records that have an empty Smart Property value. This is useful for initial backfill when you first create a Smart Property. For ongoing operations, workflow-triggered fills give you more control over when and why the AI runs.

Bulk Operations

When you first create a Smart Property, you likely need to backfill values across your existing database. Use the bulk fill option from the Smart Property settings page. Select the records you want to process (all records, a specific list, or a filtered view) and initiate the fill. Keep in mind that each record consumes 10 Breeze Credits, so a database of 5,000 contacts would require 50,000 credits for a full backfill.

Best Practices for Smart Properties

• Start with a small test batch (50 to 100 records) before running a full backfill to verify the AI produces the output you expect
• Write explicit classification instructions - vague rules produce inconsistent results
• Use dropdown field types when you need consistent categorical values rather than free-text responses
• Review AI-generated values periodically to catch drift or edge cases the rules do not cover
• Create separate Smart Properties for different classification needs rather than overloading a single property with complex multi-factor rules

Advanced Data Agent Techniques

Once you are comfortable with the basics, the Data Agent becomes significantly more powerful when combined with workflow branching, action chaining, and Breeze Intelligence data.

Branching Based on AI Output

After a Custom Prompt action that outputs a categorical value (like “High Priority” or “Low Priority”), add an if/then branch in your workflow. The HubSpot if/then branch documentation covers branching logic specifics. Route high-priority leads to immediate sales notification, medium-priority leads to a nurture sequence, and low-priority leads to a long-term drip campaign. The AI output becomes the decision point for your entire downstream workflow.

Chaining Multiple Data Agent Actions

A single workflow can include multiple Data Agent actions in sequence. For example:

1. Run a Research action to compile the account brief
2. Follow with a Custom Prompt action that reads the account brief and generates a priority classification
3. Finish with a Fill Smart Property action that stores the final classification

Each action builds on the output of the previous one, creating a layered enrichment pipeline. The key is configuring each action’s output to a specific property that the next action can reference.

Combining with Breeze Intelligence Data

If your plan includes both Breeze Intelligence and the Data Agent, you can create powerful enrichment workflows - the HubSpot Breeze Intelligence buyer intent guide covers the Intelligence side. Breeze Intelligence fills in external data (company revenue, employee count, technographic stack), and then the Data Agent analyzes that enriched data to generate classifications and scores. For example, after Intelligence populates a company’s revenue range and tech stack, a Data Agent Custom Prompt can evaluate whether the company matches your ideal customer profile based on those newly enriched fields.

Delayed Re-enrichment Workflows

Set up workflows that re-enrich records on a schedule. Use a date-based enrollment trigger (such as “Last enrichment date is more than 90 days ago”) to re-run Data Agent actions periodically. This keeps your AI-generated classifications current as contact behavior and company data evolve over time.

Error Handling and Fallbacks

Not every record will produce a clean AI output. Build fallback branches into your workflows for cases where the Data Agent returns an empty or unexpected value. Route these records to a manual review queue or a secondary workflow that attempts enrichment with a differently worded prompt.

Pricing and Plan Requirements

The Breeze Data Agent is not available on all HubSpot plans. Here is a breakdown of what each tier includes for the HubSpot Breeze Data Agent and related AI features.

Key pricing details:

• The Data Agent requires a Professional tier subscription or above
• The full Breeze AI Suite (all four agents plus Breeze Studio and Data Hub) is available at the Enterprise tier
• Each Data Agent action consumes 10 Breeze Credits per operation
• Free and Starter plans have no access to the Data Agent
• The Free CRM is limited to 2 users with limited Breeze AI features, no data enrichment, 2,000 emails per month, and basic social media scheduling only

For current pricing details, visit the HubSpot pricing page.

Frequently Asked Questions

What is the difference between Breeze Intelligence and the Breeze Data Agent?

Breeze Intelligence is a separate product that enriches your CRM with external third-party data - company revenue, employee counts, technographic information, and buyer intent signals sourced from outside your HubSpot account. The Breeze Data Agent works exclusively with data that already exists inside your CRM. It uses AI to analyze, categorize, summarize, and transform your internal records through workflow actions. Think of Intelligence as data acquisition and the Data Agent as data optimization.

How many Breeze Credits does the Data Agent consume?

Each Data Agent action (Custom Prompt, Research, or Fill Smart Property) consumes 10 Breeze Credits per operation. If a workflow contains three Data Agent actions and processes 100 records, that workflow would consume 3,000 credits total (3 actions multiplied by 100 records multiplied by 10 credits each). Monitor your credit usage in the Breeze settings dashboard to avoid unexpected depletion.

Can the Data Agent access the internet during workflow execution?

No. The Data Agent uses an offline LLM that does not have internet access during workflow execution. It can only work with data stored in your HubSpot CRM - contact properties, company records, deal information, associated records, and engagement history. If you need external data enrichment, use Breeze Intelligence as a complementary tool.

Is the Breeze Data Agent available on the Free or Starter plan?

No. The Data Agent requires a Professional tier subscription or above. The Free plan includes limited Breeze AI features but no Data Agent access, no data enrichment capabilities, a 2-user limit, and a cap of 2,000 emails per month. The Starter plan at $15 per month includes the Breeze AI Assistant and a basic Customer Agent but does not include Data Agent access. Professional Marketing starts at $890 per month ($801 per month with annual billing).

What is the character limit for Custom Prompt actions?

The Custom Prompt action has a 3,000-character limit for the prompt text. This is enough for detailed instructions with specific output formatting requirements. To make the most of this limit, be specific about which CRM properties the AI should evaluate, define the exact output format you expect, and include examples of desired responses when possible. Avoid vague instructions like “analyze this contact” and instead specify “evaluate this contact’s email engagement rate, form submissions in the last 90 days, and deal stage to classify them as High, Medium, or Low priority.”

Can I chain multiple Data Agent actions in a single workflow?

Yes. You can add multiple Data Agent actions in sequence within the same workflow. Each action’s output is stored in a designated property, which subsequent actions can reference. A common pattern is to run a Research action first to compile an account summary, then follow with a Custom Prompt action that analyzes that summary and outputs a classification, and finally use a Fill Smart Property action to store the final result. Keep in mind that each action in the chain consumes 10 credits per record processed.

Related Reading

• HubSpot Tool Page
• Best CRM Software for Small Business
• HubSpot vs Salesforce
• Best Marketing Automation Tools

Related Guides

• HubSpot Breeze Prospecting Agent: AI Sales Outreach Workflow
• HubSpot Breeze Studio: Building and Managing AI Agents
• HubSpot Breeze Intelligence: Buyer Intent Signals and Data Enrichment
• HubSpot Commerce Hub: Payment, Quotes, and Checkout Setup

External Resources

• HubSpot Breeze Data Agent Knowledge Base - Official setup, action reference, and Smart Property documentation
• HubSpot Breeze Credits Reference - Credit costs by action type and budgeting recommendations
• HubSpot Workflows API Documentation - Reference for triggering Data Agent actions programmatically