Tokens are the unit of AI consumption in ThatApp. Every interaction with AVA, every IRIS analysis, every automation that uses AI processing, and every Print document generation consumes tokens.

What consumes tokens

AVA conversations
Every message you send to AVA and every response she generates. Longer conversations with more context consume more tokens. Conversations with tool use (searches, record lookups) consume more than simple Q&A.
IRIS nightly analysis
IRIS runs every night across all connected platforms. Each analysis run consumes tokens. The more connected platforms you have, the more tokens each IRIS run consumes.
Automation processing
Automations that use AI to parse content, evaluate conditions, or generate dynamic content consume tokens when they run.
Print document generation
Documents that use AVA to populate complex variables or generate dynamic content consume a small number of tokens.

Typical usage

A single AVA conversation averages 2,000–8,000 tokens. IRIS nightly analysis averages 15,000–50,000 tokens per run, depending on the number of connected platforms. At $30/million tokens, 1 million tokens costs $0.03 per 1,000 tokens.

Most organizations with 2–5 connected platforms and moderate AVA usage consume 3–10 million tokens per month ($90–$300 in token billing).

How to see your current usage

Ask AVA: "What is my token usage this month?" She will show you:

  • Total tokens consumed this month
  • Breakdown by source (AVA conversations, IRIS, automations, Print)
  • Projected month-end total based on current pace
  • Estimated token billing at month end

If token usage is higher than expected

Common causes:

  • A team member having very long, context-heavy AVA conversations
  • An automation running much more frequently than expected (check your trigger conditions)
  • IRIS running against a very large number of connected platforms

Ask AVA: "Break down my token usage by team member and source for this month." This will show you which team member or which automated process is consuming the most tokens.

Related: How Billing Works · How to Read Your Current Usage