Skip to content
Agentguiden.no
Submit

Build a mileage log from your location history with ChatGPT

The property investor behind the Bergens Investoren account uses a private car for the company but never kept a mileage log. The investor gave ChatGPT the Google Maps location history, the property addresses and seven parking app screenshots.

Done with ChatGPTModerate

This is a public report, not advice from Skatteetaten, NAV, Altinn or an accountant. Check figures, deadlines and decisions against the original.

What was done

The property investor behind the Bergens Investoren account uses a private car for the company but never kept a mileage log. The investor gave ChatGPT the Google Maps location history, the property addresses and seven parking app screenshots. ChatGPT built a 25-page log with distances, tolls and parking, and filed the expense in PowerOffice Go. It took about 30 minutes and gave more than 9,000 kroner back. Check the original, as the investor asks accountants to weigh in.

How to do it

  1. Paste the prompt into the tool, and replace the text in brackets with your own.
  2. Read the answer through. Do not send, delete or publish anything until you have seen it yourself.
  3. Check the original before you trust a figure, a name or a date.

Try it with ChatGPT

Read my location history from [Google Maps export] for [period] and find the car trips. Match the stops to [work or client addresses], and ask me about errands you are unsure of. Build a mileage log with date, route, distance and purpose, and calculate the allowance at [rate] per km. Add tolls and parking from [receipts]. Show me the log before anything is sent to accounting. You are done when I have one PDF I can check.

Replace text in brackets with your own.

Done when. You have one PDF you can check.

Read the original

X ·

More examples

  • Ask Claude if you pay for two home insurance policies

    A homeowner had Claude, which monitors the inbox and reads a private document vault, check a renewal notice from an old insurer. Claude found two home policies on the same house for a year: the mortgage account had paid $5,292 to the old carrier and $2,883 to the new one. Claude searched 79 mailboxes, found no cancellation and drafted a backdated cancellation request. The refund is promised but not received, and the setup is homemade.

    ClaudeMy Life in the Stack on Substack ·

  • Build a negotiation sheet before bidding on a used car

    AnnaMarie Houlis at Moneywise asked Muse to prepare the purchase of a used 2004 Jeep Wrangler listed at 24,394 dollars. The agent found the price sat 3,800 to 5,900 dollars above a good deal, and that the dealer had already cut it by 1,500 dollars. It made a two-page PDF: open at 19,000, aim for 20,000 to 21,000 and walk away above 22,000. Muse made the plan, but could not negotiate with the dealer itself without more permissions.

    MuseAnnaMarie Houlis at Moneywise ·

  • Find two cheap seven-seat rental cars with Codex

    The user Xcelerate needed two rental cars with seven seats each for a trip to Maine. Codex searched in the browser through the memberships at Avis, Hertz, Costco and Chase, queued the CAPTCHAs for the user, and a sub-agent read tips on Reddit. It found an Avis reseller with a discount code at 600 dollars, against 1,100 dollars direct. The user counts it as 1,000 dollars saved, but does not say clearly whether the prices are per car.

    CodexXcelerate on Hacker News ·

Five agents a day

A short email each morning with five concrete examples of what people use AI agents for. Each example links to its source. You can unsubscribe at any time.

You will first get an email to confirm the address.