Why Was My Meta Ad Rejected? Read the Policy PDF with AI and Fix It

The rejection email tells you almost nothing

You built the ad, picked the audience, hit publish, and a few hours later Ads Manager shows a red "Rejected" label. The notice usually names a broad category, something like "Personal attributes" or "Health and wellness", and links to a policy page. It rarely says which word, image or claim caused the problem.

So you guess, change the headline, resubmit, and get rejected again. Or you open Meta's Advertising Standards and scroll through more than ten top-level sections and dozens of sub-pages.

A faster approach: treat the policy as a document. Save the relevant pages as PDFs, put your ad copy next to them, and ask specific questions until you find the rule your ad broke. Below is that workflow, plus three common rejection reasons and the questions that usually find the problem.

Step 1: Save the right policy pages as PDFs

Meta's rules live at Meta's Advertising Standards. Everything below is a paraphrase. Always check the official page before you rely on it, because Meta updates these pages without much notice.

Don't save the whole site. Start with the page your rejection notice points to, then add the pages that most often overlap with it:

1. The page named in your rejection notice.
2. Personal attributes (under Objectionable content). It catches a lot of ads that look harmless.
3. Health and wellness (under Restricted goods and services), if you sell fitness, skincare, supplements, diet or cosmetic products.
4. Prohibited commercial practices (under Deceptive content), if your ad mentions money, income or results.

To save a page as a PDF in Chrome, open it, press Ctrl+P (Cmd+P on a Mac), set the destination to "Save as PDF", and save. Expand any collapsed sections first, or the hidden text won't make it into the file.

Put the date in the file name, like "meta-personal-attributes-2026-10-12.pdf", and re-save every couple of months. Policies change, and an answer based on an old copy can be confidently wrong.

Step 2: Put your ad and the policy in one place

Create a plain text file, or a one-page PDF, with everything Meta reviewed:

• Primary text
• Headline and description
• Every line of text that appears inside the image or video (including captions and stickers)
• A one-line description of the image or each scene
• The landing page URL and its main headline
• The rejection reason, copied word for word

The text inside the creative and the landing page are the parts people forget, and reviewers look at both. Clean copy can still be rejected because the thumbnail says "Lose 10 kg in 2 weeks".

Upload the policy PDFs and your ad file together to TalkingPDF. If you haven't used a PDF chat tool before, our guide on how to chat with your PDFs using AI covers the basics. Because the answer usually depends on more than one policy page, chatting with multiple PDFs at once helps: you can ask one question and get answers that cite both the personal attributes page and the health page.

Step 3: Ask questions that point at a specific line

"Is my ad okay?" gets you a summary of the whole policy. Questions that quote your copy get you to the problem. Some that work:

• "My rejection reason is 'Personal attributes'. Quote the part of the policy that applies to the sentence: 'Struggling with your debt? This app is for you.'"
• "Which sentences in my ad file could be read as asserting or implying something about the viewer?"
• "Does the health and wellness policy say anything about before-and-after images? What conditions does it set?"
• "List every claim in my ad that promises a specific result in a specific time. Which policy section covers each one?"
• "Rewrite my headline so it describes the product instead of the viewer. Keep it under 40 characters."

Ask the AI to quote the policy text it relies on, and check that quote against your PDF. If it can't point to a passage, treat the answer as a guess.

Three common rejection reasons

1. Personal attributes

Meta's policy says ads must not assert or imply personal attributes about the person seeing them. The list on the policy page covers things like race, religion, age, sexual orientation, gender identity, disability, physical or mental health, financial status, criminal record and name.

In practice the rule is about the word "you". The ad can't suggest it knows something sensitive about the viewer. Copy like this tends to get flagged:

• "Are you over 50 and tired of joint pain?" (implies age and a health condition)
• "Struggling with debt? We can help." (implies financial status)
• "Meet other Christian singles near you." (implies religion)
• "Still anxious every morning?" (implies a mental health condition)

The usual fix moves the attribute from the viewer to the product:

• "Joint support designed for people over 50" instead of "Are you over 50?"
• "A simple plan to pay down debt" instead of "Struggling with debt?"
• "A dating community for Christian singles" instead of "Meet other Christian singles near you"

2. Before-and-after images and body image

Before-and-after images are not banned outright. Meta's health and wellness page allows some of them for certain categories, such as cosmetic procedures, when the ad is limited to adults 18 and over, and it sets conditions on weight loss content, including showing how long the results took.

What the page does prohibit is content that attacks how someone looks. Negative statements about a person's body, specific body parts or hygiene are not allowed.

So these ads usually fail for one of a few reasons: the audience includes people under 18, the copy says something like "Hate your belly fat?", the image shames a body part, or the result is shown without how long it took. Ask your PDFs "Does my copy contain any negative statement about appearance? Quote each one," then open your ad set and check the age setting. The targeting is easy to overlook when the creative gets all the attention.

3. Exaggerated claims and guaranteed results

Meta's health policy flags sensational language and promises of specific outcomes within a set timeframe without disclaimers. Its prohibited commercial practices policy covers things like deceptive pricing and guaranteed investment returns.

Copy that tends to trip this:

• "Lose 10 kg in 14 days"
• "Earn ₹1 lakh a month from your phone"
• "Guaranteed 3x returns"
• "Cures acne permanently"
• A "70% off" price that was never the real price

Usually the fix is to drop the number or describe what the product does instead of promising an outcome:

• "A 12-week plan with meals and workouts" instead of "Lose 10 kg in 14 days"
• "Learn how people run small online stores" instead of "Earn ₹1 lakh a month"
• "Helps reduce the look of breakouts" instead of "Cures acne permanently"

Ask the AI to list every number in your ad and say whether the policy treats it as a claim. This catches the ones hidden in image text and video captions.

Step 4: Fix and resubmit

Once you know which line broke which rule, make the smallest change that fixes it. Rewrite the whole ad and you won't learn anything, and you might introduce a new problem. Before resubmitting:

1. Change only the text, image or setting the policy question pointed to.
2. Re-check the landing page for the same claim.
3. Confirm age targeting if the product is in a restricted category.
4. Paste the new copy back into the chat and ask: "Does this version still conflict with the policy section you quoted earlier?"

You can make the edit in Ads Manager as usual. If you already work in an AI assistant and don't want to switch tools, you can also edit and resubmit the ad straight from Claude with a connector that links your Meta ad account and keeps every change paused until you approve it. Either way, the fix is the same: one targeted change, then re-review.

If you're sure the ad follows the policy, request another review through Account Quality instead of editing, and include the policy passage you think it complies with.

When you need to contact support

If the rejection makes no sense, or the whole account gets restricted, you'll probably end up sending screenshots to Meta support. Look at what's in the frame first. Ads Manager screenshots often show your ad account ID, billing details, other campaign names, customer names in comments, or the email address of the person logged in. Blur anything support doesn't need. A tool like Blurweb can blur personal info on the page itself before you take the screenshot, which is quicker than editing every image afterward.

Keep the chat thread from Step 3 too. If support asks why you think the ad complies, you can paste the exact passage.

An example, start to finish

Say a skincare brand runs "Tired of hiding your acne? Clear skin in 7 days, guaranteed." with a close-up before-and-after photo. It gets rejected under health and wellness. Three questions against the saved PDFs:

1. "Which sentences imply something about the viewer?" Answer: "Tired of hiding your acne?" implies a health condition, which falls under personal attributes.
2. "Which claims promise a result in a set timeframe?" Answer: "Clear skin in 7 days, guaranteed."
3. "What does the policy say about before-and-after images for this kind of product?" Answer: points to the conditions on the health and wellness page, including the 18+ requirement.

The fix: new copy reads "A gentle routine for breakout-prone skin. See how it works." The ad set is limited to 18+. The before-and-after image is replaced with a product shot. The landing page headline, which also said "7 days", is changed to match.

If you run ads often, keep a notes file next to your policy PDFs listing every rejection and the line that caused it. After a few months you'll have a list of phrases to avoid in your niche, and you'll catch most of them before Meta does.