Agents In Use
What’s new todayLast update
3 use cases added today

Grok Bot measures token usage before a workflow change

@MikeysMolecules reports using bots to measure personal token usage, then keeping Notion for project state after finding substantial usage inside individual tool loops.

Source published x.comHow we sourceAgent: Grok Bot
OUR ADAPTATION

Suggested prompt

Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.

Measure my agent usage by tool loops and starting context using the records available. Separate measurements from estimates. Recommend a project-state workflow without claiming savings we have not measured.

Evidence at a glance

Sourced account
Source poster / publisher
@MikeysMolecules
Original account
Read the source ↗
Published
Source link checked
Reported result
The poster kept Notion and introduced a read-before-work, update-after-work project-row rule. No measured token reduction after this change is reported.
Human involvement
The poster interpreted the measurements, estimated potential savings, chose Notion and says they added the fields and master view themselves.

A linked account supports attribution. It does not mean we independently reproduced its result.

Report an issue with this caseOpen correction email

Opens your email app with the case link included. Nothing is sent automatically.

How the agent did it

  1. The poster asked their bots to measure personal usage before moving project state to Grok Bot Pages.
  2. They report about 35 to 42% in tool loops, including one 73-call run using 11.1M tokens, and about 23% in starting context above a fixed floor.
  3. The poster estimated roughly 1 to 5% weekly savings from Pages with low confidence and chose to stay in Notion.

@MikeysMolecules reports having their bots measure a week of personal usage before deciding whether to move projects into Grok Bot Pages. The aim was to test a suggested workflow change against their own costs, rather than assume that another place to store project state would shorten existing chats.

The poster attributes about 35 to 42% of usage to tool loops within individual tasks. One reported run made 73 calls and used 11.1M tokens. They put about 23% in starting context above a fixed floor, and argue that Pages would only help that portion if work moved into short tasks that read the Page first.

Their estimate of roughly 1 to 5% weekly savings from switching is explicitly low confidence. It is not a measured improvement, and the capture does not supply logs or a measurement method for independent inspection.

The poster chose to keep Notion. They added Next step, Owner, Blocker, Deadline and Last updated fields, created a blocked-first master view and set a rule requiring bots to read and update the project row around each task. Those configuration decisions belong to the poster; the account does not say the bots made the Notion changes.

Prompt
Suggested prompt (written by Agents In Use, not the author's)

Source: X · @MikeysMolecules

Frequently asked questions

Was the 1 to 5% saving measured?

No. The poster labels it a low-confidence estimate of weekly savings from switching to Pages.

What did the poster measure in tool loops?

They report about 35 to 42% of personal usage in tool loops and one 73-call run consuming 11.1M tokens.

Did the bots build the Notion setup?

The source says the poster added the fields and master view. It does not attribute those changes to the bots.

More details

Remaining steps

  1. The poster added project fields and a blocked-first view, then required bots to read each project row before work and update it afterwards.

About this account

The bots reportedly measured the poster's usage. A new rule directs them to read and update project state around tasks; its compliance is not demonstrated.

Used Scout's verbatim X public API capture, read live on 2026-10-31, with explicit permission to use it instead of opening X. Reviewed the supplied post text and author bio. Media and self-thread continuations were not reviewed. Results are the poster's report, not independently reproduced. The supplied bio and post do not identify affiliation with the agent's maker or promote an owned agent product. Suggested prompt written by Agents In Use; no exact task prompt is published in the capture.

Credit: @MikeysMolecules

Source published

Added to Agents In Use

Source checked

Tools and topics

Grok Bot · Notion

Token usage · Project state · Notion

Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.