Use casesUpdated 4 min read
Research a prospect's Meta ads before the pitch
Before an agency pitches a brand, someone has to know what that brand runs on Facebook and Instagram, what its rivals keep running, and where the gap is. This page walks through that research on AdSwipeFile, step by step, and shows how to hand the legwork to Claude or ChatGPT.
Free, no card: the first 50 ads of every list, saves and 50 MCP/API calls for the life of the account for your AI agent.
What the pitch needs from ad research
A good pitch shows the prospect three things they have not put together themselves:
- What they run now. Their active ads, the formats they lean on and the hooks they repeat.
- What their rivals keep running. The ads in their niche that brands kept paying for, which is the clearest public signal of what holds up.
- The gap. A format or a hook the rivals use and the prospect does not, turned into two or three ad ideas you can show on the call.
None of this needs the prospect's ad account. It is all public, and it is the research that turns "we make great ads" into "here is what we would make for you".
The free way, in the Meta Ad Library
Search the prospect's name in the Meta Ad Library, pick its Page and read what is active. Note the start dates: an ad running for weeks has outlasted the tests next to it. Then do the same for three or four rivals.
It works, and it takes an afternoon. The library has no niche or format filter, so finding the rivals is guesswork, and Meta shows every ad while it runs. Once an ad stops, only social issue, election and political ads (kept 7 years) and ads delivered in the EU (kept a year after their last impression) stay in the library. Screenshot what you want to keep, or it may be gone by the time you build the deck.
Kept running the longest, in the library now
63,244+ ads from 1,209 brands in 26 niches. Days are how long each brand kept the ad running on Meta: a reason to study it, not a result.
Stop screenshotting ads that disappear
63,244+ winning ads, filed by brand, niche and format, and kept after Meta takes them down. Search them, or ask Claude or ChatGPT to pull them for you.
The same research on AdSwipeFile
- Open the prospect's brand page. If it is one of the brands on AdSwipeFile, its ads are filed with the longest running first. Paid plans can also find a brand in Meta's Ad Library by name and follow it.
- Open the niche. Sort its ads by longest running to see what the category keeps paying for: health and fitness ads, footwear, beauty and the rest.
- Compare formats and hooks. Each ad carries a visual format and a hook tactic. If the rivals lean on testimonial ads or a social proof hook and the prospect has none, that is your gap.
- Save the evidence. Save the ads that make your point, add a note to each, and on paid plans put them on a board.
- Share the board. A board opens read-only by link, so the prospect sees the ads before or after the call. On Agency, the shared board carries your agency's logo, and boards file under a client.
Hand the legwork to your AI agent
Some agencies that use AdSwipeFile run this research from Claude rather than the site. Connect the MCP server once, then ask: "pull the longest running ads from the prospect's three closest rivals, list their formats and hooks, and tell me which ones the prospect does not use". Your AI agent searches the library, reads the copy, and gives you the shortlist with links, ready to drop into the deck.
On a paid plan it can also save the ads it finds to a board for you, so the research and the shareable proof come out of one conversation. The Claude and ChatGPT guide has the setup.
What to say about the ads
Talk about how the ads are built and how long brands kept them, never about their results. No one outside an ad account sees spend or sales, so "this ad made money" is a claim you cannot back. "Your three closest rivals have kept a before and after ad running since spring, and you have none" is one you can.
Start before the next pitch
A free account, no card, shows the first 50 ads of every list with their copy and connects your AI agent. Pro opens every ad and adds boards and followed brands, with a 7 day free trial. Agency adds boards by client and your logo on shared boards; see pricing.
More for agencies: AdSwipeFile for agencies, the Facebook ads research tool overview, and how to track competitor Facebook ads.
Try AdSwipeFile with your assistant
Copy the setup prompt and paste it into Claude, ChatGPT or Cursor. It connects the library and asks for recent ads from your niche. Free: 50 MCP/API calls for the life of the account and the first 50 ads of every list, no card.
FAQ
How do I research a prospect's Facebook ads before a pitch?
Look up their Page in the Meta Ad Library for what runs today, then compare it with the ads their rivals kept running. AdSwipeFile files those rival ads by niche, format and hook, so the comparison takes minutes instead of an afternoon.
Can I see a prospect's ad spend or results?
No. Spend, reach and sales sit in the brand's own ad account. What you can show is how their ads are built and how long brands in their niche kept similar ads running.
How do I share ad research with a prospect?
Save the ads to a board on a paid plan and turn on its link: anyone with it sees the board read-only. On Agency, the board carries your agency's logo.
Can Claude do pitch research for me?
Yes, with AdSwipeFile's MCP server. Your AI agent searches the library, compares a prospect's ads with its rivals' and can save the shortlist to a board. Connecting is free on every plan.
What if the prospect's brand is not in the library?
Paid plans can find a brand in Meta's Ad Library by name and follow it, so its new ads arrive in your feed. Until then, its rivals' ads in the same niche still show what the category keeps running.
Sources
Checked on 6 November 2026 on each tool's own site. Prices are in US dollars as listed, and change: check the source before you buy. AdSwipeFile's facts come from its pricing page.