Boil down long documents
What it does: Hand AI a long PDF (a lease, an insurance policy, a school district letter, a 40-page report) and get the key points back in a couple of minutes, with every number and date pulled out separately so you don’t have to hunt for them. How long: 2 minutes. Does it cost money: No.
| Effort | Setup | Risk |
|---|---|---|
| ★★☆☆☆ One upload + one prompt | ★☆☆☆☆ Just the app | ★☆☆☆☆ Double-check anything you’ll act on |
How to do it
Using ChatGPT (or Google Gemini) on your phone
- Open the app, tap ＋ → File, and pick the document (PDF and Word both work; photos of printed pages work too)
- Copy and send this prompt:
Summarize this document: give me the gist in 10 bullet points or fewer. Then list separately every number, date, and dollar amount the document mentions.
- Read the summary first, then the numbers list. If a section matters to you, follow up: “Expand on point 3” or “Explain section 4 like I’m new to this”
Why the “list every number” trick matters
Ask only for a summary and AI tends to smooth over specifics: the exact rent increase, the deadline, the dollar amount. Asking for numbers and dates as a separate list forces it to surface them. That’s usually where the real content of a document lives: what changes, by how much, starting when.
Watch out
- AI summarizes, it doesn’t certify. For anything with legal or financial consequences (signing, paying, deadline decisions), use the summary to decide what to read closely, not to skip reading
- Scanned documents: if the pages are photos of paper, results get worse with bad lighting or small handwriting. Flip through what it returns against the page count; a “2-page” answer on a 30-page file means the upload failed
- Very long documents (100+ pages) may hit file size limits. Split the file or summarize chapter by chapter
Last tested: pending. This walkthrough was prepared from the products’ documented capabilities and has not yet been verified step-by-step by a real person. If you try it (whether it works or not), tell us on GitHub.