[object Object]

The Breeze pricing model rewards heavy users and surprises light experimenters. Outcome-based pricing for agents, token consumption for content generation, per-action billing for workflow runs. Without active monitoring, your monthly invoice will jump 3x in the first quarter of real usage. Here’s the playbook.

Know which Breeze action bills which way

Three pricing axes:

  • Customer Agent: per-resolution. Cheapest per-unit, scales with deflection.
  • Prospecting Agent: per-contact-touched per cycle. Mid-cost.
  • Content Agent: per-token-generated. Most variable; long-form costs add up.
  • Workflow custom code with AI calls: per-execution.

Your dashboard shows bundled “Breeze credits”; the rates differ underneath.

Set hard caps per agent

Every Breeze agent has a configurable monthly cap. If unset, the agent runs until your billing cycle ends. Set conservative caps before turning the agent on:

Customer Agent: cap = 2,000 resolutions/month
Prospecting Agent: cap = 5,000 contacts touched/month
Content Agent: cap = 200K tokens/month

Adjust upward only after observing baseline.

Build the daily usage alert

A workflow that pulls Breeze usage daily and alerts when consumption is on pace to exceed budget:

Trigger: daily 8am
Action: query Breeze usage API
Filter: projected_month_end > monthly_cap * 0.9
Action: Slack #ops-alerts with current burn rate

Audit the high-cost actions monthly

Pull the cost-by-action report. Often one or two agents account for 70%+ of spend. Question whether the value matches. A Content Agent generating 50,000 tokens per blog post is 4x what’s needed; cap the prompt length.

Distinguish trial usage from production

Marketing teams “test” content generation by burning through their test allocation in week one. Set up separate test portals or test-mode flags so the experimentation doesn’t bleed into production budget.

Don’t enable every agent at once

Roll out one agent per quarter. The first month establishes baseline cost per agent. The second month shows growth trajectory. The third validates ROI. Then you decide on the next agent. Enabling all four simultaneously means you can’t attribute cost to value.

Cancel idle agents

A Prospecting Agent enabled for a sales motion that’s been retired keeps running at low volume and accruing minimum charges. Quarterly review: if an agent’s usage is below 10% of its cap, consider disabling.

Budget for the variable, not the average

The variability across months is often 2x. A “$500/month average” Content Agent will hit $1,200 in a campaign month. Budget at peak, not average, or you’ll keep getting blindsided.

What to do this week

Pull your current month’s Breeze cost-by-action breakdown. Set monthly caps on every enabled agent. Build the daily usage alert before next week’s invoice cycle.

[object Object]
Share