Asking a PDF questions is the easy half. The form, the scan, the redaction and the pack of six files need an editor.
Upload a PDF to almost any AI assistant now and you can ask it for a summary, the key dates, or what clause 14 means. That saves reading time, but the file you get back is the file you put in.
Most of the PDF work in an office is the other half. Someone needs a field filled, a page removed, a signature added, a phone number blacked out, the words pulled off three scans. Plenty of businesses still do that half with a printer, a pen and a scanner, because the software they have will open the PDF and show it to them but won’t let them change it.
So the useful test for an “AI PDF editor” is simple. After you’ve finished, is the PDF different?
A council application, a supplier onboarding pack, a rental form. If the PDF has real form fields, open it in Oppermind’s PDF editor and fill them from one panel: text boxes, ticks, dropdowns. Add your signature and download it as an ordinary PDF that opens anywhere.
If it’s a flat PDF with no fields (common with forms that were printed and scanned once too often), add the fields yourself and type into them. You can also mark it up with highlights, comments, sticky notes or freehand drawing before you send it back.
Building a form from scratch is slow by hand. Describe the form in a sentence and the AI drafts the PDF for you. From our Academy’s own example:
“Make a fillable PDF form for new suppliers: company name, business number, bank details, a contact person and a signature field. Take the questions from the attached supplier-checklist.docx.”
Then open it in the PDF editor to add, move and test the fields and adjust the signature block before you send it. Text fields, checkboxes, radio buttons, dropdowns, date pickers and signature boxes are all there if you’d rather build it by hand.
A photographed invoice, a scanned lease, a supplier price list that only exists on paper. A scan is a picture of words, so you can’t search it or copy a number out of it.
Open the scan in the PDF editor and choose OCR (optical character recognition) from the Tools menu. It reads the pages and downloads their text as a plain .txt file you can edit, search and paste, and on a long scan it tells you if it only read the first part. The PDF itself stays a picture. OCR gives you the words, not a searchable PDF. You can also attach the scan in chat and ask for the text.
Check the numbers before you rely on them. OCR misreads smudged digits, and a 1 that became a 7 on an invoice total is exactly the mistake nobody notices until later.
A lease or a client’s invoice has other people’s details in it. How Oppermind handles the files you upload is set out in our privacy policy. Whether you may upload a particular client’s or employee’s document is a question for your own privacy obligations and client terms.
Redaction is the PDF job where getting it wrong can’t be undone, and the usual mistake is a black box.
Drawing a black rectangle over text (with a highlighter, a shape or an annotation tool) hides it on screen. The text is still in the file underneath. Select the black bar, copy, paste into a blank document, and there it is. The ABA Journal walked through exactly this in January 2019, after lawyers for Paul Manafort filed a document in a US court whose “redacted” passages could be read by highlighting the black bars and pasting the text somewhere else.
Australian courts warn about the same thing. The Federal Court of Australia’s guide to redacting electronic documents tells people not to use the annotation tool to draw black boxes over text, because the text can still be found by searching or by copy and paste. It also says text turned white in a word processor stays readable once copied. The Victorian Legal Practitioners’ Liability Committee gives the same warning to lawyers, and tells them to remove metadata from a document before it goes out. General information only, not legal advice. If you’re filing with a court, follow that court’s current practice notes.
In Oppermind’s PDF editor you switch on Redact and drag across the text you want gone. It works on text you can select, so a scan with no text layer, or words inside a picture, can’t be redacted this way. When you choose Download PDF, the editor doesn’t lay boxes over the original. It builds a new PDF from an image of each page with the redacted areas filled solid black, and it carries none of the old file’s hidden text, metadata, form fields, attachments or edit history across.
The trade-off is that the redacted copy is image-only. You can’t select or search any of its text, and form fields are flattened. Your original stays editable in the workspace, so you redact a copy, not your only version. Only Download PDF makes the redacted copy. While a file has redactions waiting, the editor switches off Convert to Doc and its other text exports, and it makes page images and password-protected copies from the redacted version. The file you send is still the downloaded PDF.
Whatever tool you use, check the file before it leaves the building. Open the downloaded copy and zoom in on each black bar to make sure no edge of a letter shows around it. Then run the test the ABA Journal recommends. Select the area where the text was, copy, and paste it into a blank document. If nothing comes through, it’s gone.
The quote, the terms and three certificates, as one file. Merge them in the editor, and split a long file back into sections when a client only needs part of it.
For the attachment that’s too big to email, choose Compress from the Tools menu. The editor tells you the size before and after, and leaves the file alone if it can’t make it smaller. For a pack with personal details in it, Password protect downloads an encrypted copy and leaves the file you have open unchanged. Send the password another way (a text message, a phone call), not in the same email. You can also ask for either one in chat.
Sometimes the right move is to stop editing the PDF. Convert to Doc pulls the text out of the PDF into a document you can rewrite in full.
Know what you get. The converter takes the PDF’s text, page by page, and drops the layout, so columns, tables and images need rebuilding. A scan has no text to pull, so run OCR on it instead and paste the text into a new document. For a two-page letter that’s fine. For a designed 40-page brochure, ask the AI to rebuild it as a new document or PDF from a one-line brief instead.
| Job | What a chat-about-your-PDF answer gives you | What the job needs | Where it is in Oppermind |
|---|---|---|---|
| Form sent to you | A list of the fields | Fields filled and signed in the file | Fill form fields, Sign |
| Form you send out | A list of questions | A fillable PDF | A one-prompt draft, then the PDF editor’s form fields |
| Scan | The text, sometimes | Editable text you can check | OCR in the Tools menu (a .txt download), or ask in chat |
| Redaction | Nothing, or a summary without the details | The details removed from the file | Redact, then Download PDF for the image-only copy |
| Pack of files | A summary of each | One file, small enough to send | Merge, split and Compress in the editor, Password protect for the copy you send |
| Back to editable | The text pasted in chat | A document you can rewrite | Convert to Doc |
If you already pay for Adobe Acrobat, you’re paying for a mature PDF editor, and Acrobat does most of these jobs well. Adobe’s Australian pricing page (accessed 24 September 2026) lists Acrobat Standard at A$23.99 a month, Acrobat Pro at A$31.99 and Acrobat Studio at A$39.99, each on an annual plan billed monthly, including GST (month-to-month is A$39.99, A$47.99 and A$55.99). On that page, redaction and turning scans into searchable PDFs are listed under Pro and Studio, and the AI features (asking AI to edit, convert or compress PDFs, and getting summaries and answers) are listed under Studio.
Oppermind is not affiliated with, endorsed by, or sponsored by Adobe. Adobe and Acrobat are trademarks of Adobe Inc. Prices and plan contents change, so check Adobe’s page before you decide.
Oppermind works differently in a few places. It doesn’t edit a PDF’s existing text in place. The redacted copy is image-only. OCR gives you a text file, not a searchable PDF. And OCR, Compress and Password protect take files up to 22 MB. If your work depends on Acrobat-specific features, compare them on Adobe’s own page first.
Oppermind is an all-in-one AI workspace with its own editors, and the PDF editor is one of them. The AI makes a PDF from a sentence, reads one you attach, and handles OCR, compression and passwords when you ask in chat. The editor is where you change the file by hand: fill, sign, redact, merge, compress, lock and convert.
The same workspace makes the document the PDF started as, the spreadsheet the scanned invoice belongs in, and the deck the report turns into, so a PDF job rarely ends at the PDF. Plans are Free with no card, Starter A$9.95, Pro A$29.99 and Pro Plus A$59.99 a month (prices at 24 September 2026).
Whatever you’re here to make, make more of it.
Both, in different places. The AI reads and summarises PDFs, drafts new ones from a description, and in chat can run OCR on a scan, compress a file or add a password. Changes to an existing file (filling, signing, redacting, merging) happen in the PDF editor, where you can see every change before you download. The editor’s Tools menu has OCR, Compress and Password protect as well.
No. Download PDF rebuilds the file from page images with the redacted areas filled in, so the original text isn’t underneath. Zoom in on the black bars and do the copy-and-paste test on anything sensitive anyway.
Yes. Run OCR from the PDF editor’s Tools menu and the text downloads as a .txt file, or attach the scan in chat and ask for the text. The PDF itself isn’t made searchable. Check figures against the scan, because OCR can misread smudged characters.
For the office PDF jobs in this post (forms, signatures, redaction, merging, compression, scans, conversion), it can stand in, inside a workspace that also makes documents, sheets and decks. It works differently in places. It doesn’t edit a PDF’s existing text in place, the redacted copy is image-only, and OCR gives you a text file rather than a searchable PDF. If your work depends on Acrobat-specific features, compare them on Adobe’s own page first.
A full PDF editor next to the AI, with documents, sheets and decks in the same workspace.