🏷️ Robot shoppers.
Shopping agents get increasingly involved in your checkout process, you’ll be able to create Google AI campaigns by describing goals, and more.
In this issue
Good morning.
Okay, books are getting an “organic” label now. A UK startup is stamping human-written titles with a little thumbprint, certifying them as made by real people. No machines… or additives.
We can see this catching on. Organic emails. Free-range LinkedIn posts. Hand-picked birthday texts from your aunt. Next up: a stamp proving a human actually read the terms and conditions.
Never mind. Some things are too far-fetched.
AI agents can now check out. Is your product data ready?

Remember when Alexa was going to do all our shopping? Mostly it set kitchen timers.
Now AI assistants can search your catalog, fill a cart and pay.
Agents got their own checkout lane at Shopify. Shop MCP lets any compatible agent search 1B+ Catalog products, build carts and complete Shop Pay checkout.
Those orders land in your Shopify admin like any other. The agent only knows what your catalog tells it, though: price, stock, and product details
Still, a big shift in online commerce. And if you need another sign that it’s moving quick…
Card networks want proof, too. Amex’s new framework records the merchant, budget and product preferences, then locks them into a Proof of Intent Token.
For example, if you ask for white Nikes in size 9, Amex can block an agent that tries buying Adidas. That means that fuzzy specs make your match harder to prove.
Budget and deals do the filtering on AI search surfaces too:
- ChatGPT is testing a shopping box where users pick a product type and budget, then hit “Find” to refine the query.
- Google’s AI Mode Shopping ads now show offers like “Buy 4 get $20 off”, with a details popup, the same as on the regular results page.
Even pharma got a seat. Google is testing branded prescription drug ads in AI Mode and AI Overviews, US only.
Few categories come with more ad rules than pharma, so Google clearly sees AI Mode as a real ad channel.
Whoever ends up doing the buying, your product data is the first thing they read.
Campaign setup by prompt hits testing as AI Max adds a matching report
Mad Libs handed you a page of blanks to fill.
Google Ads is testing the reverse: you write one sentence about your goal, and its AI fills in the rest.
Describe it, and Google builds it. Some advertisers now see a campaign builder that starts with “What’s your goal for this campaign?”
Google’s AI suggests goals and assets based on the answer.
Ginny Marvin, Google’s Ads Liaison, confirmed it’s an experiment. You can still click “Create campaign manually.”
Lookalikes get looser. Demand Gen Lookalike segments are moving to suggestion mode by default. That means they no longer work as a strict limit on who sees your ads.
To keep the old behavior, fill out Google’s opt-out form. There’s no date yet.
Finally, AI Max shows its matches. The new Search terms and ad combinations report shows each search term next to the headlines, descriptions, and landing page it served. It also shows:
- If the match came from AI Max or URL inclusions
- If you provided the assets
Check your top-spending terms first. Search Engine Land’s example: a “basic car wash” search served with a $699 ceramic coating ad.
More API updates. Starting with v26, the Google Ads API releases twice a month. There are still four major versions a year, and the extra minor releases don’t break anything.
Sunset dates stay the same, so you won’t have to migrate more often.
Google wants AI to make more setup and targeting calls. At least you’re getting more ways to check them.
84% of marketing leaders expect a budget squeeze, and over half are stuck in reactive mode

CMOs are on the hook for growth, yet many control it about as much as a Love Island contestant controls the next twist.
The Awin US CMO Study 2026, conducted with Business of Marketing, puts numbers on it.
More than half are playing catch-up, with 51% in Reactive mode. Only 2.5% make the report’s top Resilient tier.
Two channels do the heavy lifting for most teams:
- 70% get at least half of new customers from their top two channels
- Only 2% have a diversified mix
New customers find you elsewhere, too. 53% say at least half first hear about their brand through creators, publishers, marketplaces, communities or AI answers.
Teams with a structured partner program do better: 61% are ahead of target, vs. 49% of everyone else.
The report scores teams on measurement, speed, channel mix and more, so you can see where you stand.
Your AI visibility score is a coin toss. Stop flipping it once…

Ask AI Mode the same question three times and you’ll rarely see the same answer twice.
So much for that screenshot in the monthly report.
Greg Jarboe argues SEOs now work in the news business, and he lays out how to measure AI visibility that never sits still.
One prompt, many universes: SE Ranking ran 10,000 keywords through AI Mode three times on the same day. Average URL overlap was 9.2%, and 21.2% of keywords had none.
SparkToro and Gumshoe put the odds of getting the same brand list twice at under one in 100. A University of St. Gallen study saw about 65% of cited sources change daily.
One check tells you what happened once.
Platforms move the goalposts, too: Promptwatch saw Reddit's share of ChatGPT Search citations drop from 3.83% to 0.52% within weeks.
Ahrefs found AI Overviews and AI Mode share only 13.7% of their citations.
Track each platform separately, because a blended average buries a sharp ChatGPT break. Cross-check sudden drops with a second source.
Even Promptwatch called its Reddit figure provisional.
Run it until it means something: SparkToro points to 60 to 100 runs per prompt. St. Gallen suggests seven or more daily runs per prompt, over rolling two to four week windows.
Report a rate with a range, and track brand presence over individual URLs, since it holds steadier.
Keep a diary, not a dashboard: Log model releases and platform updates next to your own publishing and PR.
- Visibility up everywhere after a big data release, competitor flat? That's a signal.
- It jumps on one platform just after a retrieval change? Thank the platform.
Act like a newsroom: Ahrefs found AI-cited URLs average 1,064 days old, versus 1,432 for organic results. Muck Rack found earned media drives 84% of AI citations, while owned media manages just 13.7%.
Publish original data worth covering, refresh revenue pages when the substance changes, and get your PR team pitching.
AI visibility behaves like weather. Nobody judges the climate by one rainy Tuesday.
How 30k+ marketers catch social trends before they peak

Your brand doesn't need another 'very demure, very mindful' post three weeks too late.
Steal a move from Netflix, Apple, and Disney: their social teams read Geekout, which serves up:
- The latest social media news
- Expert insights you can actually use
- Hit videos and memes before your feed catches up
Join 30k+ social media managers and marketers who'd rather start trends than chase them.
Hiring managers have a favorite two-letter word
Can you guess which one? No, you don’t have to guess twice.
If your résumé doesn’t mention AI yet, hiring managers are noticing.
CXL scanned 2,126 LinkedIn marketing job posts from Q1 to Q3 2026, and AI went from a nice-to-have to a requirement fast.

Top findings:
- AI now appears in more than half (54%) of all listings.
- Jobs that require AI nearly tripled in two quarters (15% → 40%).
- Mentions of AIO or GEO have more than doubled in Q3 alone (6% → 15%).
One caveat: Q3 covers only US jobs, so the numbers aren’t a perfect comparison. Still, every category grew every quarter.
Companies have moved past testing AI and now want people who can run it every day. The jump in AEO/GEO shows that AI search has become a real line item in marketing budgets.
Writing “familiar with AI” in your CV is not enough. Show results.
Add specific workflows, automations, or AI search wins to your portfolio, and use the same keywords that show up in job posts.
AI MARKETING: This daily newsletter condenses the latest and greatest AI developments into a 5-minute read. They read the noise, pull out the signal, and explain what actually matters for work, strategy, and the tools your team is already using. Trusted by 700k+ readers. Subscribe right here.*
FACEBOOK: Users can now tell Feed and Reels what they want to see, in their own words, through a new “Your algorithm” section in the three-dot menu. After a “Not interested” tap, follow-up questions may ask why, and your feed adjusts right away.
META: Engagement, Leads, App promotion, and Sales campaigns no longer let you pick placements at the ad set level. Awareness and Traffic keep manual control. To block a placement, use account-level exclusions. Ads also won’t run where your media doesn’t fit, so add more formats, up to 10 per ad.
YOUTUBE: A limited test brings Ask YouTube to some Shorts. Viewers tap a sparkle icon and ask questions about the video, in English and only if they’re 18 or older. Over 140 million people used it on the regular watch page in June, so expect viewers to ask about your Shorts too.
BING: Topics Portfolio, a beta addition to the AI Performance report in Webmaster Tools, groups queries where AI answers cite your site into topics, with total citations and the number of sites cited for each. Another view lists queries where competitors get cited, and you don’t. Interesting.
*This is a sponsored post.
Get the next issue in your inbox.
Delivered Monday, Wednesday, and Friday, our carefully curated digital marketing news, tech, and actionable advice can be consumed in 7 minutes or less.



























