Key Takeaways
- Assume nothing behind a form will ever be read. An assistant does not enter an email address to download your guide. Every gated file is invisible by design, which is a choice worth keeping and worth knowing about.
- Check whether your PDF contains text or pictures of text. Open it and try to select a sentence with your cursor. If nothing highlights, the words are images and no assistant is reading them either.
- Treat slide decks as the worst format for being quoted. Slides carry fragments rather than sentences, and most get published as images. There is rarely a full claim in there for anything to lift.
- Publish the argument as a web page and keep the file as the deliverable. The page is what gets read and cited. The download stays exactly where it is, still doing its job of collecting addresses.
- Link every file you want found from a page that people actually reach. An unlinked file sitting in a media folder is reachable in theory and undiscovered in practice, because nothing points a crawler toward it.
A consultant spends three weeks on a forty-page methodology guide, puts it behind an email form, and gets six hundred downloads. It is the best thing they have ever written about their own work. No AI assistant has ever mentioned it, and none ever will.
The usual explanation offered is that the document is too long, or too detailed, or too commercial. None of those is the reason.
The reason is more boring and much easier to fix. Most of that material never reaches a point where anything could read it in the first place.
Do AI assistants read PDFs, slides, and lead magnets?
A public PDF that is linked from somewhere and contains real text can be read and quoted, much like a web page. A slide deck usually cannot, because its words are pictures. A lead magnet behind an email form is never read at all, because nothing fills in your form.
So the answer splits three ways, and the split has nothing to do with quality. It is about whether the words are fetchable, extractable, and findable, in that order.
That is worth knowing before you spend another month on a document. The format decides whether the argument inside it can ever travel.
The three gates every file has to pass
Three things have to be true before a sentence you wrote can appear in somebody else's answer. Miss any one of them and the other two stop mattering.
Reachable. The file has to be fetchable without a form, a password, or a login, and it must not be blocked from crawling. Google's own introduction to robots.txt describes the file that quietly excludes whole folders, and a surprising number of sites block their own downloads folder without ever meaning to.
Readable. The words have to exist as text. Google publishes a list of the file types it can index, and PDF is on it, alongside common word processor and presentation formats. What is not on any such list is a picture of a page. A scanned document is an image, and an image of a sentence is not a sentence.
Referenced. Something has to link to it. A file uploaded to a media folder and never linked from any page is technically public and practically undiscoverable, because nothing leads anywhere near it.
An ordinary web page
Clears all three. This is the format everything else is measured against.
A PDF with real text in it
Clears all three when it is public and linked. Google lists PDF among the file types it indexes.
A PDF that is scanned pictures of pages
Fetched, then unreadable. The words are pixels, not text.
A slide deck exported as images
Same failure. Your headline is a picture of a headline.
A lead magnet behind an email form
Stopped at the first gate. Nothing fills in your form.
A file nobody links to
Readable and reachable, and still never found, because no path leads to it.
Only two of the six reach the far side. Both of those are the same thing in different clothing: public text that something links to.
The lead magnet row is the one that surprises people. It fails at the first gate, before readability is even a question, and it fails for a reason you deliberately built.
What happens to each format
The differences are practical rather than technical. Here is what actually becomes of each thing you might have published.
| What you published | Can it be fetched | Can the words be read | What to do |
|---|---|---|---|
| An ordinary page on your site | Yes, if it is linked and not blocked | Yes | Nothing. This is the format to copy. |
| A PDF with selectable text | Yes, if public and linked | Yes | Keep it, and put its argument on a page as well |
| A scanned or image-only PDF | Yes | No, the words are pixels | Republish the text, or retype the key claims onto a page |
| A slide deck exported as images | Yes | No, and slides carry fragments anyway | Write the talk up as prose; the deck was never the argument |
| A lead magnet behind an email form | No | Not reached | Keep the gate, and publish a public summary in front of it |
| A file uploaded but never linked | In theory only | Yes, if anything found it | Link it from a page a person would actually visit |
| A video of you explaining it | Yes | Only through captions or a transcript | Publish the transcript as text on the same page |
Two rows in that table are the same problem in different costumes. The scanned PDF and the image-exported deck both contain your best sentences, and both store them in a form that cannot be quoted.
Why your best material is your least visible
Because you gated it on purpose, and the gate works. Most people put their strongest thinking behind the form precisely because it is strong enough to be worth an email address, which means the material most likely to earn a mention is the material nothing can see.
The honest version of this: do not tear the gate down. It is collecting addresses, and that is a real business function that being quoted does not replace. Anyone telling you to un-gate everything is trading a working system for a hope.
What you can do is stop treating the gated file as the only place the argument lives. The guide can stay behind the form while the three claims that make it good sit on a public page, in text, where they can travel. The same reasoning applies to results locked in a case study document, which is covered in making client results legible to AI.
One limit worth stating plainly. Publishing the summary does not guarantee anything quotes it. It moves the material from a place where it definitely cannot be read to a place where it can, and that is the whole of what this step buys you.
What to publish instead of moving the file
Write one page that states what the document argues, in complete sentences, with the specifics left in. Not a teaser and not a table of contents. The actual claims, worded so that somebody reading only that page learns something true.
Give it the shape that gets lifted easily. Google's guidance on creating helpful content keeps returning to the same idea: content written to answer a real question beats content arranged to rank. A page with a direct question as a heading and a direct answer beneath it is the most liftable thing you can build, and the reason tables and lists work so well is covered in why AI loves tables and lists.
Then link the download from that page. The page earns the mention, the file collects the address, and neither one has to pretend to be the other. Which kinds of pages tend to get quoted is worked through in what kinds of pages AI assistants quote most.
If you want to know whether any of your material is being reached today before you rewrite anything, the free check asks ChatGPT 25 real buyer questions from your niche at $0 and emails a plain-English readout. It runs as soon as you ask for it, and the report lands within the hour between 7am and 5pm Central.
Questions we hear the most
Do AI assistants read PDFs?
Sometimes. A PDF that is public, linked from somewhere, and contains real selectable text can be fetched and quoted like a web page. A scanned PDF, or one locked behind a form, is not read at all.
Can ChatGPT read my slide deck?
Rarely, and usually not in a useful way. Most decks are published as images, so the words are pictures rather than text, and slides carry fragments instead of the complete sentences an answer needs to quote.
Will AI read a lead magnet that sits behind an email form?
No. Nothing fills in your form, so the file is never fetched. That gate works exactly as intended for lead capture, and it also makes the material inside completely invisible.
Should I delete my PDFs and put everything on web pages instead?
No. Keep the files, because they do a job that pages do not. Add a public page that states the same argument in text, and link the download from it.
How do I tell whether my PDF has real text in it?
Open it and try to select a sentence with your cursor. If the text highlights, it is readable. If nothing highlights, the page is an image and the words cannot be extracted.
How do I find out whether AI mentions my business at all?
Ask ChatGPT the questions your buyers would ask, in their words rather than yours, and write down who gets named. Do it in a fresh logged out session so your own history does not skew the answer.
What does the AnswerHalo free check include?
It asks ChatGPT 25 real buyer questions from your niche and emails you a plain-English report: your score, who got named in your place, and the one fix to start with. It costs $0.
How quickly does the free check come back?
It runs as soon as you ask for it, and the report is in your inbox within the hour, 7am to 5pm Central, or by 10am the next morning. There is no call and no credit card.
Your buyers are already asking. Find out what AI tells them.
The free check asks ChatGPT 25 real buyer questions about your niche and sends you the report within the hour, at $0. If it shows you are already getting named everywhere, we will say so in plain words.
Get my free check →