How this helps you
You see exactly which routines and triggers spend your usage, with a ranked list of safe changes you approve one by one.
Run a read-only usage efficiency audit across my entire Grok Bot team. 1. Routine inventory: ask every Bot, including hidden Bots, to report its active routines, schedules, triggers, purpose and recent runs. Tell me anything you cannot inspect. 2. Frequency: flag hourly routines, short intervals and jobs that run more often than their results are needed. Recommend longer intervals or on-demand runs. 3. Triggers: find routines watching broad events such as every Slack message or every email, and recommend specific channels, keywords or conditions. 4. Duplicate work: find Bots checking the same sources or producing overlapping reports, and recommend one owner. 5. Task scope: find recurring jobs with vague or broad instructions and recommend tighter inputs, clear completion conditions and shorter outputs. 6. Workflow: find repeat tasks that could use an existing skill or connected app instead of navigating websites. Don't claim savings without measurements. 7. Usage review: check the usage information you can see, separate estimates from measured usage, and rank changes by expected impact. Give me one report: routines worth pausing, schedules worth reducing, triggers worth narrowing, duplicate work worth combining, workflows worth simplifying, and your top [five] changes, with evidence and trade-offs for each. Do not pause, delete, edit or create any routines, Bots, skills or connections without my approval.
Fill in: [five]: how many top changes you want ranked. Everything else works as written; it's a shortened version of Michael's prompt with the same seven checks.
What they did
Michael Fenech read the official Grok Bot docs to work out what actually spends usage. The finding that stood out: routines. The routines help page says each run spends usage, and an hourly schedule, a short interval or a Slack trigger on a busy channel can use a week of usage in a day. It's not how many messages you send, it's how much work your Bots do. So he wrote a 7-point audit to paste into a chief of staff Bot: list every routine on every Bot (hidden Bots too), flag schedules that run more often than the results are needed, find triggers that watch broad events like every Slack message or every email, spot two Bots doing the same job, tighten vague recurring instructions, find website clicking that an existing skill or connected app could replace, and rank changes by expected impact while keeping estimates apart from measured numbers. It ends with one report and a hard rule: nothing gets paused, deleted, edited or created without his approval. His advice: start with the routines that run most often. You don't necessarily need fewer Bots, just Bots that only do necessary work.
Who it's for: Anyone with several Bots and routines whose weekly usage runs out early.
Before you start
- A chief of staff or main Bot that can see your other Bots.
- A few routines already running (otherwise there's nothing to audit).
How to try it
- Paste the prompt below into your chief of staff Bot as one message.
- Read the report starting with the routines that run most often; those are where a change saves the most.
- Check the evidence it gives for each item. If it says it couldn't inspect a Bot, open that Bot's Details tab and look at its routines yourself.
- Approve changes one at a time by replying with exactly which routine to pause, slow down or narrow.
- Run it again after your next weekly reset and compare the report with the first one.
What to expect
One report listing, for example, a 15-minute inbox check that found new mail in only a few runs (suggest twice a day), a Slack trigger on a busy channel (suggest one channel and one keyword), and two Bots both building a morning summary (suggest one owner), each with its evidence.
Common mistakes
- Letting it fix things straight away. The approval line at the end is what keeps a cleanup from switching off a routine you rely on.
- Treating estimates as facts. Michael asks it to keep estimated and measured usage apart; trust the measured ones first.
- Cutting Bots instead of work. His point is that a Bot that only runs when needed costs little.
- Forgetting hidden Bots. Ask for them by name in the inventory, or they can keep running unnoticed.
Related tips
- Make routines cheaper: events, slow schedules, change checks: A 5-minute routine wakes 8,640 times a month. Use an event trigger, the slowest schedule that works, and a cheap change check.
- Build a monthly usage audit page that labels every routine: Hand your Bot screenshots of your usage and billing pages, get a page that marks each routine Keep, Slow down, Change check or Retire.
- Out of usage? Run these five checks before you buy credits: Grok Bot says you're out of usage but you still have credit? Check the weekly pool, the monthly cap, the meter, the account, then what's eating it.
Worth knowingParaphrased from Michael Fenech's post on X; the prompt is a shortened version of his, with the same seven checks and approval rule. The usage line is from the official routines help page.
Learn more: Grok Bot Routines Not Working? Do It by Hand, Save a Skill, Then Schedule · All Grok Bot routines tips