Your competitors' newsletter ads, read for you every Monday (with Claude and SponsorGap)

Somewhere this week, a competitor of yours tried something new. Maybe a free trial instead of the usual discount, a newsletter they've never sponsored, or a third booking in the same newsletter, which usually means it's working for them.

You won't see it. It ran in an issue you don't subscribe to, between two stories you'd never read.

A post about having Claude listen to 500 podcasts for a business made the rounds recently. Its point stuck with us: someone has already solved your problem, and they're explaining it 34 minutes into an episode you'll never hear. Newsletter and podcast ads work the same way. Your competitors' tests, offers and wins are published every week. Nobody has time to read them all.

SponsorGap already reads them. We scan newsletters, podcasts and YouTube, find the sponsors, and save the ads. This month we added the part that was missing: deciding which of those ads matter to you.

What you get every Monday

Tell SponsorGap who your competitors are, and optionally what you're working on. Every Monday you get a brief with the 3–5 moves that matter this week. We look for specific signals in our data:

• A new placement. A competitor ran in a newsletter for the first time. That's usually a test.
• A rebooking. A competitor booked the same newsletter again. Brands don't pay twice for something that doesn't work.
• A new offer, angle or promo code. For example, a competitor that always ran "20% off" switched to "start your free trial".
• A competitor that stopped. They ran in a newsletter regularly, then went quiet for a month while the newsletter kept running other sponsors. That's an opening.
• A brand new to your industry. New money in your market, and often a future competitor.

Our AI then reads these against your profile and keeps the few that matter. Each comes with:

• What we saw: the brand, the newsletter, the date, the audience.
• The ad itself, word for word where we have it, plus the promo code and a link to the issue.
• Why it matters: one or two sentences, written for your business.
• Do this week: one concrete next step.

Here's what one item looks like. The brands are made up; the format is real:

Facts come from data, not from the AI

The podcast post was honest about what went wrong: the AI got facts wrong that were sitting in the company's own notes, and machine transcripts made their own mistakes. The fix was a second pass that checked every claim, and keeping "what the speaker said" apart from "what we concluded".

We built that split into the brief from the start:

• The AI can't invent a placement. It can only choose from the moves we found in our data. The brand, newsletter, date, ad text and link always come from our records. The AI writes only the "why it matters" and "do this week" lines.
• Ad copy is checked. When we save an ad word for word, we confirm the text actually appears in the issue. If our AI paraphrased it, we drop the copy and keep the link instead.
• "Stopped" means stopped. We only flag a competitor leaving a newsletter if that newsletter still ran other sponsors, so a gap in our own scanning never looks like a competitor walking away.
• Our read is labelled as ours. Every brief says it plainly: what we saw is from our data; "why it matters" is our read.

It learns what you care about

Every item has Useful and Not relevant buttons, in the email and in Marketing OS. Next week's brief takes them into account. Your profile also has two optional fields:

• What you're working on, e.g. "Launching our new plan to mid-size SaaS teams this quarter; testing newsletters before podcasts."
• Leave out of the brief: brands or newsletters you never want to see.

The brief is only as good as what it knows about you, so one honest sentence about your goals does more than any setting.

Or ask Claude directly

The brief is also available in Claude, through the SponsorGap MCP connector. Three tools matter here:

get_competitor_changes deliberately leaves out our AI's opinion. In your Claude chat, Claude already knows things we don't: the deal you're closing, the launch next month, the channel your CFO wants to cut. So we hand it the facts and let it judge, the same way the podcast setup worked: our data as the input, your context as the filter.

Prompts to try:

• "What did my competitors do in newsletters this week? Which of these should I act on, given we're launching our free tier in December?"
• "Read every ad Northwind ran in the last 90 days. How do they usually pitch, and what changed recently?"
• "Show my weekly brief, then draft a newsletter ad that answers Northwind's new free-trial offer."
• "Which newsletters did a competitor stop sponsoring this month? Rank them by fit for us."

How to set it up

1. In SponsorGap, open Marketing OS, add your website, and confirm your industry and competitors. Add a sentence about what you're working on.
2. Your first brief arrives on Monday. It's in your inbox and on the Your market page.
3. To use it in Claude, get your API key in SponsorGap under Settings → API Access, then follow the MCP setup guide to add SponsorGap as a connector in Claude (it also works in Cursor and other MCP clients).

The AI picks, the ad copy and the feedback loop are part of SponsorGap Pro, and MCP access is on Professional plans and above. Free accounts get a shorter Monday brief: the top two competitor moves and a count of what else changed.

What it can't do (yet)

• We only see what we scan. Coverage is broad across newsletters, podcasts and YouTube, but not complete. If a competitor only advertises in a newsletter we don't track, we won't see it. Tell us which ones you'd like added.
• Word-for-word ad copy is new. We save it for ads found since November 2026. Older ads have a summary, a short quote and a link to the issue.
• No click data. We can't see whether an ad converted. A rebooking is the best signal we have that something works, which is why we flag it.

The bottom line

The useful insight in that podcast post wasn't really about podcasts. It was this: the agents are only as good as what they know about you, and the input is everywhere if something reads it all.

For sponsorships, the input is your competitors' ads. SponsorGap reads them. Claude, or our brief, tells you which ones matter. You spend five minutes on Monday instead of never knowing.

Set up your competitors in Marketing OS →