When should you use it?
- You see duplicate charges or billings after you canceled
- You have receipts and cancel confirmations to support refund requests
- You want a tracker that separates requested, approved and received refunds
Not for: Letting the agent send disputes or share card details without your OK; Inventing hardship stories or entitlements the evidence does not support; Treating refund speed stories in the source post as guaranteed outcomes
What do you need, and what do you get?
You need
- Receipts, renewal notices and proof of any cancellations
- Transaction dates, amounts and merchant names from your statements
- Any deadlines or terms pages you already have
You get
- A list of suspect charges with evidence and likely refund path
- Draft requests grounded in transaction details, held until you approve send
- A tracker showing which refunds you asked for, which were approved and which have arrived
How does it work, step by step?
- Gather evidencePoint the agent at receipts, renewal emails, cancel confirmations and matching statement lines.
- Classify chargesHave it flag duplicates, post-cancel billings and other refund candidates, and ask about anything unclear.
- Prepare requestsCheck terms and deadlines, then draft each request from real transaction details only.
- Approve, send, trackApprove each send yourself. Track requested, approved and received amounts separately.
The prompt
The original prompt by @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, ChatGPT agent and Claude. Values in lime are placeholders.
Original prompt by @Michael_Fenech_
Review my recent receipts, renewal emails and cancellation confirmations for duplicate charges, charges after cancellation and purchases that may qualify for a refund. Ask me about anything unclear. Check the relevant terms and deadlines, then prepare each request using the transaction details and supporting evidence. Don’t invent a reason or claim an entitlement you can’t establish. Get approval before sending. Track each approved request through to a response, and distinguish refunds requested, approved and actually received.
Exactly as published by @Michael_Fenech_ in X post: 4 money-saving audits (subscriptions, internet/phone, insurance, unwanted charges).
Our adapted version (optional)
Fill in the blanks (optional; the copied prompt uses your values)
Audit recent charges for refund candidates. Evidence I am providing: {receipts_and_emails}. Statement lines: {statement_lines}. 1) List suspect items: duplicates, charges after cancellation, and other purchases that may qualify for a refund. Ask me about anything unclear. 2) For each candidate, note transaction details, supporting evidence, relevant terms or deadlines, and a plain refund path. 3) Draft each request using only those details. Do not invent a reason or claim an entitlement you cannot establish from evidence. 4) Wait for my approval before sending any request or opening a dispute. 5) After I approve, track each item as requested, approved or actually received. Keep those three statuses distinct. Never send mail, chat or card-network disputes on your own.
What can go wrong?
- Do not invent a reason or claim a right the terms do not support.
- Never send a dispute or share full card numbers without your approval.
- A promised refund is not money received; keep those statuses apart.
- Deadlines matter; a late request may need a different path.
Our take
Refund work fails when the ask is vague or the follow-up disappears. Fenech's prompt insists on evidence, forbids invented entitlements, gates every send and splits requested, approved and received. That tracking line is the part most people skip, and it is why promised credits get forgotten. Distinct from a single unauthorized-renewal email playbook, this is a multi-charge audit across receipts, renewal mail and cancel confirmations. Deadlines in the merchant terms decide whether you still have a clean path or need a different route. Never invent a hardship story, and never let the agent open a card dispute or share full card numbers alone. We leave the post's two-hour refund and $70 recovery as speaker-claimed. In our dry run with fictional receipts (one duplicate streaming charge, one gym billing after cancel confirmation, one unclear marketplace purchase), the agent asked about the marketplace order, drafted two evidence-backed requests with deadlines cited from the terms we provided, waited for send approval, and tracked one refund received and one still only requested.
Where the time goes back: Suspect charges are batched with evidence and deadlines in one pass, so you send fewer weak requests and can see which refunds actually landed.
FAQ
How is this different from unauthorized-renewal-refund-emails?
Will the agent email merchants for me?
Are the refund stories in the X post guaranteed?
Source and credit
The original prompt on this page is by @Michael_Fenech_, shown word for word with a link to where it was published. The summary, notes and adapted version are ours.
- Original
- X post: 4 money-saving audits (subscriptions, internet/phone, insurance, unwanted charges)
- Creator
- Michael Fenech (@Michael_Fenech_)
- Published on
- X
- Original date
- Oct 23, 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
- Oct 23, 2026: passed. Dry-run tested by our AI test agent on sample input. Evidence-only drafts; send hold; requested vs received tracked.
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
- Oct 23, 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.