Token Budget & AI Usage

Overview

Every AI interaction in MileIntel — concierge questions, agent runs, and analysis — consumes tokens. Your plan includes a monthly token budget, and MileIntel tracks your usage in real time so you always know where you stand.

How It Works

1. Budget Allocation — Each plan tier has a monthly AI budget:

- Traveler — $0.50 (enough for basic concierge questions)

- Navigator — $5.00 (supports regular concierge use + several agents)

- Executive — $20.00 (heavy concierge use + many agents)

- Black — $50.00 (unlimited practical usage)

1. Status Bar — The budget indicator appears in the bottom status bar across the app. It shows a progress bar with your current spend and remaining budget.
2. Usage Tracking — Each concierge message and agent run logs its token cost. You can see the per-interaction cost in the Activity Feed.
3. Budget Enforcement — When you hit your monthly cap, AI features pause until the next billing cycle. You'll get a warning at 80% usage.

Key Details

• Shared agent analysis (routes watched by multiple users) splits the token cost across all watchers, making it much cheaper per person.
• Agent runs use the efficient Haiku model via Batch API, which costs roughly 50% less than interactive queries.
• The budget resets on your billing date each month, not on a calendar month boundary.
• Token costs are measured in dollars for simplicity, but they represent actual AI compute usage.

Tips

• If you're on a lower tier and running out of budget, reduce agent deployments first — they run repeatedly and add up faster than occasional concierge questions.
• Check the per-agent token usage in Agent Workstations to identify which agents are the most expensive.
• Upgrading your plan is the simplest way to increase your budget if you're consistently hitting the cap.