Breeze AI Onboarding.
One agent, done right.
Most teams are kicking the tires on Breeze — sidebar questions, a few generated emails, a look at the agent marketplace. That's curiosity, not deployment. The add-on builds one Breeze agent end to end — prerequisites, guardrails, audit discipline — live with your team, plus a working session on Breeze Assistant for daily use. Diamond Partner, strategy-led, done-with-you.
Breeze AI Onboarding is a Layer 03 service.
Every service we offer maps to one of four layers. Breeze AI Onboarding sits in the Kernel — the layer that owns revops & hubspot admin.
Deployment, not curiosity.
Breeze isn't one thing — it's three connected layers. Breeze Assistant is the in-app helper, included with seats. Breeze Agents are autonomous workers installed from the marketplace and billed by credits or outcomes — Prospecting works buying signals, Customer answers across channels grounded in your knowledge base, Content writes in your brand voice, and the Knowledge Base, Social, Personalization, Data, and Studio agents each have their own job. Breeze Intelligence sits underneath with enrichment and buyer intent.
The catch: an agent is only as good as the CRM data, brand voice, and knowledge sources behind it. Turning it on is the easy part. Configuring it so it produces work your team trusts — with governance your team can audit and guardrails that prevent embarrassing failures — is where most teams stall. The prerequisites are the actual work, and most teams skip them.
So we deploy one. One agent right beats five agents wrong: the pattern installed during this engagement — prerequisites first, guardrails reviewed, audit cards read, credits managed — becomes the template for every agent after it. And the team learns to work with the agent in live sessions, not just observe it: when to trust it, when to override, when to escalate.
A clean kernel produces useful agent output; a messy kernel produces output nobody trusts — which is why Breeze AI pairs naturally with the hub onboardings and why custom MCP servers extend the same pattern beyond Breeze when the business requires it.
The process.
Strategy · 1 session
Recommendations, not how-to. Breeze readiness and use-case assessment — which single agent delivers the most value first. Data and brand-voice prerequisites. Governance, guardrails, human-in-the-loop, and audit-card review. Credit and outcome-based cost modeling, so usage isn't a surprise. A roadmap: one agent now, the rest sequenced. You keep the advisory artifact.
Build for you · off-call, async
Heads-down, off-call: brand voice trained correctly, knowledge sources loaded (knowledge base, site, docs), data hygiene for the chosen agent, initial agent configuration connected to the right channels and objects, and test scenarios prepared for the workshops.
Build with you · 3–4 workshops
Half the engagement, deliberately. The chosen agent configured live — Customer Agent grounded in the KB, Prospecting Agent on buying signals, Content Agent in brand voice. Responses and guardrails tuned, audit cards reviewed, Run Agent triggers set where relevant. Test, QA, and go live together — plus a working session on Breeze Assistant for the team's daily work.
Training · 1 session · recorded
Role-based. Working with the live agent: monitoring, overrides, handoff to humans. Reading audit cards and managing credits and outcomes. Broader Breeze Assistant usage across roles. SOP and recording handoff. Additional training rolls into the retainer.
What "done" looks like.
The engagement runs 4–6 weeks across 5–6 meetings. Build-for-you is async, so live time goes to strategy, workshops, and training.
Things people ask before booking.
Is this required, like the hub onboardings?
No — Breeze AI is an add-on, available alongside any hub onboarding or on its own. HubSpot doesn't require it. Most teams pair it with the hub that matches the agent: Service Hub with the Customer Agent, Sales Hub with Prospecting, Content Hub with Content. The kernel work and the agent work coordinate, which is why the combined engagement runs efficiently.
Can't we just turn an agent on ourselves?
You can — that's exactly how most stalled Breeze deployments start. The agent is the visible part; the brand voice, data hygiene, knowledge structure, governance posture, and audit-card discipline behind it are the work. Skip the prerequisites and the agent produces output nobody trusts. We do the prerequisites correctly first, then configure the agent live so your team learns to operate it, not just observe it.
What's not included?
One agent is built and configured at this price — additional agents are add-on scope or retainer work; we build one agent end to end rather than three poorly. Also carved out: agent usage fees (credits or outcome-based — a client cost, billed by HubSpot; we model the expected usage); large-scale Data Agent engagements against complex architectures; net-new brand voice development and knowledge-base content writing (Voice services); custom MCP servers beyond Breeze; and training beyond the session caps, which rolls into the retainer.
What do we save by bundling or committing?
Bundled with one or more hub onboardings, the foundation work runs once — two or more programs take 10% off the stacked total, and the full suite takes 20%. Signed alongside a 12-month $2,000/mo fractional admin retainer, the add-on takes 20% off: from $3,500 to $2,800 — and the retainer keeps the agent tuned after launch.
Which agent should we deploy first?
The strategy phase formalizes it, but the common patterns: Customer Agent for teams with support volume and an existing KB (fastest ROI). Prospecting Agent for a working Sales Hub with buying-signal data. Content or Social Agent for teams with an established brand voice. Knowledge Base Agent when knowledge is scattered. Data Agent when data quality is the bottleneck blocking everything else. Studio when your workflow doesn't fit any marketplace agent.
See how breeze ai onboarding fits into The Kernel.
Which agent pays back first is a question for the Diagnostic. Ninety minutes at the kernel whiteboard answers it — you leave knowing whether the Breeze AI add-on fits now, and which agent to start with if it does.