← Back to library

Audit how you manage your agent team, not only the agents

Michael Fenech (@Michael_Fenech_) argues the next gain for a multi-agent setup is often how the owner manages the team. His eight-point audit looks at delegation, clarity, interruptions, corrections, bot selection, decision bottlenecks, workload and follow-through. We explain how to use it and show his prompt word for word with credit.

  • OutcomeHave your chief-of-staff agent review the last 30 days of how you delegated, clarified, interrupted, corrected, assigned, approved and followed up across your agent team, score eight management areas with evidence, list your three biggest weaknesses, and draft a seven-day improvement plan, without changing any bot, routine or instructions until you approve.
  • SetupUnder an hour
  • LevelIntermediate
  • CreditMichael Fenech (@Michael_Fenech_)
View original
Works with
Grok BotClaude CoworkChatGPT agentCustom agent

Published Nov 6, 2026Updated Nov 6, 2026AI dry run on sample input: passed Nov 6, 2026Source verified Nov 8, 2026

When should you use it?

  • You already have several agents and feel like the bottleneck is you, not them
  • The same corrections keep coming up and you suspect the instructions are not the only problem
  • Approvals pile up while some agents sit idle

Not for: A setup with one agent you check every day; Platforms where a lead agent cannot see other agents' conversations or handoffs; Replacing a tool or skill audit; run those separately when the agents themselves are the issue

What do you need, and what do you get?

You need

  • A lead or chief-of-staff agent that can read recent conversations, assignments, corrections and approvals
  • About 30 days of team activity it can access
  • Your willingness to read hard feedback without asking it to rewrite bots yet

You get

  • Numeric scores for each of the eight management areas, with two examples and one improvement each
  • Your top three management gaps, separated into observed versus suspected
  • A seven-day improvement plan that changes your behavior first
  • No bot, routine or instruction changes until you approve them

How does it work, step by step?

  1. Paste it into your lead agentUse the chief of staff or whichever agent can see the others' work.
  2. Let it gather evidenceAllow it to ask other bots for examples when your own history is thin.
  3. Read observed versus suspectedTreat unverified areas as gaps in access, not as free passes.
  4. Pick one weakness for the weekStart the seven-day plan with the change that removes the biggest bottleneck.
  5. Approve before any bot editsOnly after the plan works on your side should you change prompts, routines or assignments.

The prompt

The original prompt by Michael Fenech (@Michael_Fenech_), word for word, with a link to where it was published. Below it is our adapted version with fill-in blanks, if you want it. Works in Grok Bot, Claude Cowork, ChatGPT agent and others. Values in lime are placeholders.

Original prompt by Michael Fenech (@Michael_Fenech_)

Word for wordby Michael Fenech (@Michael_Fenech_)
Run a MANAGEMENT OPTIMISATION AUDIT across my Grok Bot team.

Analyse how I've managed my Bots over the last 30 days.

Review the conversations, task assignments, corrections, approvals and handoffs you can access. 

Ask other Bots for relevant evidence where needed.

Evaluate my management behaviour across these 8 areas:

1: DELEGATION — Am I giving Bots clear outcomes, or micromanaging every step?

2: CLARITY — How often do my instructions lack context, success criteria or clear boundaries?

3: INTERRUPTIONS — Am I changing priorities or redirecting Bots before they finish?

4: CORRECTIONS — Am I repeatedly correcting the same mistakes instead of fixing the underlying instructions?

5: BOT SELECTION — Am I assigning work to the right Bots, or creating unnecessary handoffs?

6: DECISION-MAKING — Am I becoming a bottleneck by requiring approval for low-risk decisions?

7: WORKLOAD — Am I overloading certain Bots while others are underused or duplicating work?

8: FOLLOW-THROUGH — How often do I assign work without checking the result or closing the loop?

For each area, provide:

• A score out of 10, based on available evidence.
• Two specific examples of my behaviour.
• The impact on team performance.
• One practical improvement I can make.

Then identify my three biggest management weaknesses and create a 7-day improvement plan.

Be direct. Don't flatter me or invent evidence.

Separate what you observed from what you suspect. If you can't access enough history to assess an area, mark it unverified rather than guessing.

Don't change any Bots, routines or instructions without my approval.

Exactly as published by Michael Fenech (@Michael_Fenech_) in Management optimisation audit for a Grok Bot team.

Our adapted version (optional)

Fill in the blanks (optional; the copied prompt uses your values)

Adapted prompt1 variable
Act as my chief of staff and audit how I have directed the agent team for the past {days} days. Read-only.

Inspect recent threads, task briefs, fix-ups, sign-offs and transfers you can see. Ask sibling agents for examples when your own view is thin.

Rate these eight themes from 1 to 10. For each, give two concrete examples, the effect on delivery, and one change I should make: handing off outcomes, briefing quality, cutting in mid-task, repeating the same fix-ups, picking the right agent, approval bottlenecks, uneven load, and closing the loop.

Finish with my top three management gaps and a one-week plan that changes my habits first.

Stay blunt. No praise and no invented proof. Label gaps you could not verify. Change no agents, schedules or instruction files until I say so.

What can go wrong?

  • A flattering audit is useless. Keep the rule that it must not invent evidence or soften the scores.
  • Fixing a bot when the real issue is unclear briefs will waste the week. Change your instructions to agents first.
  • Low-risk approvals you insist on signing yourself are a common bottleneck. Name which decisions can clear without you.
  • Interrupting mid-task trains agents to wait for the next redirect. Finish or formally re-brief instead.

Our take

Most agent-team advice obsesses over prompts and tools. Fenech's point is sharper: a polished roster still stalls when the owner micromanages, redirects mid-task or never closes the loop. The eight areas cover the real failure modes we see in multi-agent setups, especially repeated corrections that should have become instruction fixes and approval gates on low-risk work. We also like the hard ban on inventing evidence and on changing bots without approval, which keeps the audit diagnostic instead of destructive. British spelling in his original prompt is left as he wrote it; our page uses US spelling around it. Our caution: the audit is only as strong as the history the lead agent can see, so unverified areas need more access, not optimistic scores. In our dry run on a fictional 6-bot consulting team, it scored decision-making and interruptions lowest, cited concrete redirect examples, and proposed a seven-day plan that pre-cleared three low-risk actions and banned mid-task rebriefs.

Where the time goes back: You stop spending cycles rewriting agent prompts when the slowdown is unclear briefs, duplicate assignments or approvals only you can give, and you leave with a one-week plan aimed at your own habits.

FAQ

How is this different from a skill or tool audit?
Those inspect what the agents do. This one inspects how you assign, correct and approve their work.
What if it cannot see enough history?
It should mark that area unverified. Widen access or paste examples rather than letting it guess.
Will it work outside Grok Bot?
Yes, anywhere a lead agent can review handoffs and approvals across a small team of agents or assistants.

Source and credit

The original prompt on this page is by Michael Fenech (@Michael_Fenech_), shown word for word with a link to where it was published. The summary, notes and adapted version are ours.

Original
Management optimisation audit for a Grok Bot team
Creator
Michael Fenech (@Michael_Fenech_)
Published on
X
Original date
Nov 6, 2026
License
Not openly licensed; the original prompt is shown word for word with credit and a link to the original
Originally for
Grok Bot
AI dry run
Nov 6, 2026: passed. Dry-run tested by our AI test agent on sample input. Evidence-based scores, observed vs unverified separated, plan targeted owner behavior, no unauthorized changes.
AI dry run on sample data; not run with live connected accounts. The playbooks.bot test agent filled the prompt with synthetic inputs, produced the output and checked it against the page's stated outputs and guardrails.
Verified
Nov 8, 2026: source link resolves to the original, the credited creator published it, our excerpt or prompt matches it, and the license allows our use.

Are you the creator? Claim or correct this page.

More in Operations & Admin

All 203 Operations & Admin playbooks →

Related playbooks

Operations & AdminPrompt✓ AI dry run

Brief an agent with nine fields, not a how-to

Before you hand a job to an agent, fill in nine short fields that name the result, the sources it may use, the deadline, what it may do alone, what needs your say-so, what it must never do, how to prove the work is finished, what to do when it gets stuck, and where to put the deliverable, then add a line that tells it to pick the simplest reliable method.

Credit: Michael Fenech (@Michael_Fenech_)

Minutes
Operations & AdminPrompt✓ AI dry run

Brief your agent on the outcome, not every step

Hand a new job to your lead agent by describing the finished result instead of the steps, and have it confirm what success looks like, ask only for context it cannot find itself, pick the simplest dependable approach, state what it will do alone and what needs your approval, and check the result against your goal before calling it done.

Credit: Michael Fenech (@Michael_Fenech_)

Minutes
Operations & AdminPrompt✓ AI dry run

Build a dashboard page your agent refreshes on approval

Have your agent interview you about what a dashboard should track, where each number comes from and how fresh it must be, then build one page with the most important items on top, a last-checked time on every section, clear flags for missing or stale data, and a proposed refresh schedule that waits for your approval.

Credit: Michael Fenech (@Michael_Fenech_)

Under an hour
Operations & AdminPrompt✓ AI dry run

Check which of your agents should be merged or retired

Have your lead agent map every agent on your team with its jobs, routines, skills, connections and recent activity, find overlap and unused output, recommend keep, absorb, retire or hand back to you for each, and draft a safe wind-down plan covering routines, open work, saved instructions, files and sign-ins, with nothing removed until you approve.

Credit: Michael Fenech (@Michael_Fenech_)

Under an hour
Operations & AdminPrompt✓ AI dry run

Find routines your agents run twice and merge them

Have one agent collect every scheduled routine across your agent team, compare the saved instructions of any that look alike, flag pairs that read identical sources or run too close together, and hand you a short keep, merge or cut table with the expected saving, without changing anything until you say yes.

Credit: Michael Fenech (@Michael_Fenech_)

Minutes