What Breeze Agents Actually Are

Breeze Agents are HubSpot's autonomous AI workers. Unlike standard HubSpot AI automation workflows, agents make judgment calls. Unlike Breeze Copilot (which assists you in real time) or Breeze Intelligence (which enriches your data), Agents operate independently. You configure them, set guardrails, and they execute—prospecting for leads, drafting content, scheduling social posts, and resolving support tickets without waiting for a human to press a button.

That autonomy is what makes them powerful and what makes setup critical. A well-configured agent is a force multiplier. A poorly configured one is a machine that generates bad outreach, off-brand content, or incorrect support responses at scale.

This guide covers all four agents with the operational detail your team needs to get them running correctly. For the broader Breeze AI landscape—including Copilot and Intelligence—read the full tactical breakdown.

The Four Breeze Agents at a Glance

Each agent requires a Pro or Enterprise subscription to its respective Hub. Starter and free tiers don't have access. If you're evaluating whether to upgrade, factor agent capabilities into your AI ROI measurement calculation—they often justify the tier jump on their own.

Pricing: The Outcome-Based Model

HubSpot moved Breeze Agents to outcome-based pricing in 2026. Instead of paying per seat or per message, you pay for results.

The outcome-based model changes the economics entirely. You're not paying for AI to try—you're paying when it delivers. At $1 per qualified lead, the Prospecting Agent needs to generate only a handful of meetings per month to outperform the cost of a part-time SDR. At $0.50 per resolved conversation, the Customer Agent pays for itself the first week it deflects routine tickets from your support queue.

Credits and pricing details vary by contract and tier. Confirm your specific allocation with your HubSpot account team before budgeting.

AI Readiness Assessment: What Your Portal Needs Before Activation

Every Breeze Agent shares the same foundational requirement: clean, structured data. Agents amplify whatever is in your CRM—good or bad. If your data is messy, agents scale the mess.

Universal Prerequisites

• HubSpot Pro or Enterprise on the relevant Hub
• Contact duplicate rate under 5%. Agents can't distinguish between duplicate records. Two records for the same person means duplicate outreach or conflicting support responses.
• Email validity rate above 90%. The Prospecting Agent sends emails to addresses in your database. Invalid addresses mean bounces, which damage your domain reputation for all emails—AI and human.
• Required fields populated on 80%+ of records. Agents use contact and company properties to personalize output. Incomplete records produce generic, low-performing results.
• Consistent property naming conventions. Agents rely on property values to segment and target. If "Industry" has 47 variations of "Software," targeting breaks down.

If your portal doesn't meet these thresholds, pause on Agents and run the HubSpot data quality AI checklist first. Two to four weeks of cleanup will dramatically improve agent performance from day one.

Setting Up the Prospecting Agent

The Prospecting Agent is the most impactful—and the most sensitive to configuration quality. Get this one right and it fills your pipeline. Get it wrong and it burns through your prospect list with bad outreach.

Step 1: Navigate to Agent Settings

Go to Settings → AI → Breeze Agents. Select the Prospecting Agent and toggle it on. HubSpot walks you through initial configuration, but the defaults aren't optimized for B2B mid-market—customize every step.

Step 2: Define Your Target Audience

Set your ICP criteria: target industries, company size ranges, job titles, geographic focus, and any custom properties that define your ideal buyer. Be specific. Broad targeting produces broad (read: generic) outreach.

Use your existing closed-won deal data to reverse-engineer criteria. What do your best clients have in common? Start there.

Step 3: Configure Email Templates and Brand Voice

Upload or select your best-performing email templates. The agent uses these as structural and tonal references. If your existing sequences have strong open and reply rates, feed those to the agent first.

Set explicit brand voice guidelines: tone, terminology to use, terminology to avoid, and any compliance language required in outbound communications.

Step 4: Set Sending Rules and Cadence

Define sending windows (business hours in your target's time zone), daily send limits per contact, and sequence length. Start conservative—lower volume with higher quality beats mass-blasting your prospect list on day one.

Step 5: Configure the Approval Workflow

For the first two to four weeks, require human approval on every agent-generated email before it sends. Review the quality, personalization accuracy, and tone. Once you're confident in the output, move to a sampling model—approve 1 in 10 instead of every message.

Step 6: Monitor and Optimize

Track open rates, reply rates, and qualified lead conversion weekly. If performance drops, check targeting criteria and template quality first. The agent learns from engagement signals, so early performance data shapes long-term results.

Setting Up the Content Agent

The Content Agent generates blog posts, landing page copy, case studies, and marketing emails. It's a production accelerator—not a replacement for your content strategy.

Configuration Essentials

• Set your brand voice profile. Upload your style guide, sample content pieces, and terminology preferences. The more reference material you provide, the more on-brand the output.
• Define content categories and topics. Map your pillar-spoke content AI mesh architecture so the agent generates within your strategic framework rather than producing random topics.
• Establish the review workflow. Every piece of agent-generated content must go through human review before publishing. No exceptions. AI writes competent drafts. Humans add the voice, insight, and strategic nuance that make content worth reading.
• Set SEO parameters. Input target keywords, meta description guidelines, and internal linking rules so the agent generates content that's optimized from the first draft.

The Content Agent is most effective when you treat it like a junior writer who's fast and tireless but needs editorial direction. Give it a detailed brief and it delivers a solid first draft. Give it a vague prompt and it delivers generic filler.

Setting Up the Social Agent

The Social Agent maintains your company's social media presence—generating posts, scheduling them across platforms, and adapting based on engagement data.

Configuration Essentials

• Connect your social accounts. Link LinkedIn, Facebook, Instagram, and X accounts through HubSpot's social tools.
• Set posting frequency and timing. Define how many posts per week per platform and preferred publishing windows based on your audience's active hours.
• Provide content pillars. Tell the agent which themes to post about: product updates, industry insights, team culture, client success stories. Without guardrails, it defaults to generic engagement-bait.
• Enable approval for the first two weeks. Review every post before it goes live. Adjust tone and content direction based on what the agent produces, then relax to a sampling review model.

The Social Agent won't make you a thought leader. It will keep your company page active and consistent, which is the baseline most B2B companies struggle to maintain. Think of it as removing a recurring operational burden rather than as a growth lever.

Setting Up the Customer Agent

The Customer Agent resolves tier-one support conversations using your knowledge base, help docs, and historical conversation data. It's the fastest path to measurable support cost reduction.

Configuration Essentials

• Build or audit your knowledge base. The agent is only as good as the content it draws from. If your help docs are outdated, incomplete, or poorly organized, the agent gives bad answers. Invest in your knowledge base before activating.
• Define resolution criteria. What constitutes a "resolved" conversation? Set clear criteria so the agent doesn't prematurely close tickets that need human follow-up.
• Configure escalation triggers. Define when the agent hands off to a human: sentiment detection (frustrated language), topic complexity (billing disputes, contract questions), or VIP account flags.
• Set the confidence threshold. The agent should only respond autonomously when it has high confidence in the answer. Low-confidence queries should route to your team. Start with a high threshold and lower it as the agent proves reliable.
• Monitor resolution quality weekly. Sample resolved conversations to verify accuracy. A resolved ticket with the wrong answer is worse than an unresolved ticket—it erodes client trust.

The Customer Agent delivers the fastest ROI of all four agents when your knowledge base is strong. Teams with comprehensive help docs see 30–50% ticket deflection within the first month.

Best Practices Across All Agents

Pair agent output with HubSpot AI lead scoring to prioritize which leads agents engage first.

• Start with approval workflows on everything. No agent should operate fully autonomously on day one. Review, calibrate, then gradually loosen the reins.
• Treat the first 30 days as a calibration period. Agent output improves as it processes more data and feedback. Early performance is not final performance.
• Clean your data first. This cannot be overstated. Every agent problem we've diagnosed in client portals traces back to data quality. It's always the data.
• Set explicit guardrails. Define what agents should never do: contact competitors, send after-hours emails, respond to legal inquiries, or engage with accounts flagged as do-not-contact.
• Measure outcomes, not activity. Don't celebrate that the Prospecting Agent sent 500 emails. Celebrate that it generated 12 qualified meetings. Output-based pricing reinforces this mindset—lean into it.

For a deeper dive on where AI agents should operate versus where humans should stay in control, read our AI vs. manual processes comparison.

Frequently Asked Questions

How do I set up Breeze Agents in HubSpot?

Navigate to Settings → AI → Breeze Agents in your HubSpot portal. Select the agent you want to activate and toggle it on. Each agent requires specific configuration: the Prospecting Agent needs ICP targeting criteria and email templates, the Content Agent needs brand voice guidelines and content briefs, the Social Agent needs connected accounts and posting schedules, and the Customer Agent needs a populated knowledge base and escalation rules. Start with approval workflows enabled so you can review agent output before it reaches your audience, then loosen controls after a two-to-four-week calibration period.

What Breeze Agents are available in HubSpot?

HubSpot offers four Breeze Agents: the Prospecting Agent (identifies and contacts potential leads through personalized outreach sequences), the Content Agent (generates blog posts, landing pages, case studies, and marketing emails), the Social Agent (creates and schedules social media posts across connected platforms), and the Customer Agent (resolves tier-one support conversations using your knowledge base). Each agent requires a Pro or Enterprise subscription to its respective Hub—Sales Hub for Prospecting, Marketing Hub for Content and Social, and Service Hub for Customer.

How much do HubSpot Breeze Agents cost?

Breeze Agents use outcome-based pricing. The Prospecting Agent costs approximately $1 per qualified lead—you pay when a contact meets your qualification criteria and responds to agent outreach. The Customer Agent costs approximately $0.50 per resolved conversation—you pay when the agent resolves a support inquiry without human escalation. The Content Agent and Social Agent are included in Marketing Hub Pro and Enterprise subscriptions with usage credits based on your tier. Exact pricing and credit allocations vary by contract, so confirm your specific terms with your HubSpot account team.

What data quality is needed before activating Breeze Agents?

Your portal should meet these baseline thresholds: contact duplicate rate under 5%, email validity rate above 90%, required fields populated on at least 80% of records, and consistent naming conventions across all properties. Agents amplify whatever is in your CRM. Clean data produces targeted, personalized agent output. Dirty data produces duplicate outreach, bounced emails, and generic messaging at scale. Most portals need two to four weeks of focused data cleanup before agents deliver reliable results. Use the data quality checklist to assess your readiness.

Can Breeze Agents replace my sales and support teams?

No. Breeze Agents handle specific, repeatable tasks within your revenue operation—they don't replace the humans who run it. The Prospecting Agent scales top-of-funnel outreach but can't negotiate deals or build relationships. The Customer Agent deflects routine support tickets but can't handle complex escalations or emotionally charged situations. The most effective approach is treating agents as specialized teammates who handle volume work, freeing your human team to focus on high-judgment, high-stakes activities that drive revenue. Read the full Breeze AI tactical breakdown for the complete capability map.

Ready to Activate Your Agents?

Setting up Breeze Agents is straightforward. Setting them up so they actually perform takes operational expertise—clean data, precise targeting, calibrated guardrails, and a clear measurement framework.

Get an AI readiness assessment and we'll evaluate your portal's data quality, process maturity, and agent-readiness so you launch with confidence instead of crossed fingers.

Want to explore on your own first? Start with Mission Control on Launchpad for self-guided resources on preparing your HubSpot portal for AI.