Skip to main content

Agent Actions: How Agent Usage Is Measured

How Weflow counts Agent Builder usage in agent actions, what one action covers, your monthly allowance and what happens when it runs out.

Agent Builder usage is counted in agent actions. One action is roughly one million tokens of AI processing, so an agent run that consumes three million tokens counts as three actions.

Your workspace has a monthly allowance set by its plan, and when it runs out, agents stop rather than running into overage.

What an agent action is in Weflow

An agent action is counted from the tokens a run consumes, at roughly one million tokens per action. A scheduled run uses at least one action, even an agent with no AI Agent step.

Test runs use no agent actions; Weflow limits test runs that go beyond fair use, at its own discretion.

Weflow does not sort agents into fixed types with fixed prices: the count follows the work each run does.

Agent actions are the only metered part of Weflow's AI.

Note: Apart from agent actions Weflow's AI features are included in your pricing and don't consume agent actions:

  • Ask Weflow AI

  • Meeting summaries, AI Field Updates, meeting prep and coaching scorecards

  • Agents you build in Claude, ChatGPT or your own platform on Weflow's public API and MCP connector

How many agent actions an agent run uses

The count depends on how much data a run reads and reasons over:

  • About one action: a simple scheduled agent, such as a weekly summary of a small set of deals

  • Several actions, typically up to five: an agent that reasons across many call transcripts, emails and activities. When the data is more than one model can hold at once, the AI Agent step splits the job across several agents, and each part adds tokens

A daily digest of deals that changed stage is cheap enough to run every day. An analysis of every conversation in the pipeline against a methodology uses far more, so schedule it with care.

In the flow editor, the Estimated cost icon gives the cost in actions per workflow run.

Because actions vary with the data an agent ingests and processes, Weflow can only estimate: run a test, review the usage after the first full run, and keep reviewing it regularly.

Agent action allowances by plan

Tier

Agent actions per month

Price

Free

25

Included in every Weflow plan

Growth

500

$299 per month

Scale

2,500

$999 per month

Enterprise

Custom volume

Custom package

The allowance is per workspace, not per user, and usage resets on the 1st of each month. Admins can build an unlimited number of agents on every tier; only Weflow admins can build, edit and manage agents.

To move to a larger tier, click Get More Actions on the Agent Usage page.

What happens when agent actions run out

When your workspace has used its actions for the month, Weflow stops its agents until actions are available again. It never bills overage and never moves your workspace to a larger tier by itself. The agents restart at the monthly reset, or as soon as you move to a larger tier.

Because agents stop rather than overrun, you can start on the included allowance, watch how many actions your agents use, and size the tier from real usage.

How to see agent action usage

  • Agent Usage page: in Admin Console > Weflow AI > Agent Usage, see your plan's monthly amount, the renewal date, this month's usage against the previous month, a chart of actions over time and Usage by Workflow: each agent's runs, total actions, average actions per run and share of the total

  • Usage summary: on the AI Agents page, the Usage summary button shows the actions used against your plan and the reset date, with View Usage leading to the Agent Usage page

  • History: an agent's View History lists every run with its Consumed Actions, and can be filtered to test runs

Agent Usage page in the Admin Console: the monthly Agent Actions plan, renewal date, Get More Actions, the usage summary and actions over time

Frequently asked questions about agent actions

Does Ask Weflow AI use agent actions?

No. Ask Weflow AI comes with every Weflow plan and is not metered.

Do agents I build with the MCP connector use agent actions?

No. Agents you build in Claude, ChatGPT or a data warehouse on Weflow's API and MCP connector use no agent actions. Only runs inside Agent Builder are metered.

When does agent action usage reset?

On the 1st of each month. The Usage summary and the Agent Usage page show the reset date and the days left.

Is there a limit on how many agents I can build?

No. Every plan can build an unlimited number of agents; only the actions their runs use count against the allowance.

Did this answer your question?