Almost Friday, folks.
If a publisher has pitched you an AI visibility package this quarter, you're not the only one.
The Daily Mail is running 12 to 15 GEO tests for brands across the US, the UK and Australia, and plans to turn them into a product early next year. Its director of SEO and editorial ecommerce, Carly Steven, said so at Digiday's Media Strategies London event, per Digiday.
The method: work out the prompts consumers use when researching a brand's category, track them across AI platforms "over a very long period of time," and see which sources get cited. Where publishers get cited and the brand is weak, the team writes bespoke branded content.
For one unnamed personal finance brand, visibility across the target prompts went from about 50% to nearly 75%, she said.
She's open about the motive, calling it "quite an exciting new potential revenue opportunity – not just a defensive play." Future's GEO product has more than 30 clients. Ziff Davis has several. Forbes is building a consultative offer around its Vetted reviews.
So AI visibility is turning into a media buy, and the seller is the publisher the model already cites. Fine by me. Paying a publisher to write about you is older than search.
The measurement part is where I'd slow down. "There’s a lot of awareness that there is a gap in the attribution model," Steven said. Her best example: one brand found Claude was 0.1% of referrals in its analytics, but around 11% of conversions once it asked customers on a form how they'd heard of it.
That's one brand, told by the seller. But a form field asking customers how they heard of you costs nothing.
Before you buy, ask for the prompt set, the baseline, how long they tracked it, and which cited pages they wrote. Then add the form question and check their numbers against your own customers.
If a publisher offered you 75% visibility tomorrow, how would you know it worked?
Today’s sponsor:
Leave Granola and get up to 12 months free of Wispr Flow Notetaker + Dictation
If you have paid time left on an individual Granola plan, we'll match it with a Wispr Flow subscription that includes Notetaker and dictation, and add bonus time, up to 12 months total. Sign in or create a Wispr account and submit proof of your plan to check eligibility.
TALK ON THE MARKET
Kevin Indig, author of Growth Memo, on testing whether AI visibility shows up later:
Rare to see someone publish the test that didn't work. It's why I'd want the form question before the invoice.
Want to see case studies of AI-powered campaigns? We publish a daily edition on the most creative use cases of AI in advertising.
Interested in AI innovation across Europe? We publish a weekly round-up on AI-Ready Europe:
NUMBERS MATTER
0.62%. ChatGPT sessions as a share of Google organic sessions, across 12 months of GA4 data from 100 sites and more than 145 million sessions, in Marie Haynes' analysis published October 21. ChatGPT visitors converted better, at 13.8%, but Google organic drove 27x more conversions than all LLMs combined.
One consultant's sample, and referral data undercounts AI. Still, size the GEO line to match.
DAILY PROMPT
Our Survey Question Generator, for the one question your analytics can't answer:
Design a customer research survey for the objective below. Research any relevant benchmarks or question-wording conventions on the web if that improves the instrument, and read any attached files (past survey results, churn data, call transcripts, ICP notes) before drafting.
## Research Details
- **Objective**: [Describe what you want to learn from customers, e.g., "Understand why customers churn after 6 months"]
- **Target Audience**: [Who you're surveying, e.g., "Mid-market SaaS customers with 50-500 employees"]
- **Survey Length**: [Number of questions, e.g., "8-10 questions"]
- **Response Format**: [Choose: Multiple choice, Likert scale, Open-ended, or Mixed]
## Requirements
- Clear, unbiased wording: no leading, double-barreled, or loaded questions
- Mix of quantitative and qualitative items so results can be counted and quoted
- Logical flow from easy screening questions toward sensitive ones
- Language matching how this audience actually talks, jargon only where they use it
- Skip logic wherever it reduces respondent effort
## Output Format
Provide questions numbered 1-[X], with:
- The exact question text
- Response options (if applicable)
- Brief note on what insight this question provides
- Any skip logic recommendations
Then add a short section naming the two questions most likely to produce biased or unusable data, what you would change, and which analysis each question supports.Set the objective to where buyers first heard of you, AI assistants included. Then put the answers next to your referral report.
CMO CORNER
Leaders who rely on visibility often lack alignment.
A visibility score is easy to report and easy to buy. The harder part is getting sales, finance and you to agree on what it's supposed to move.
We provide consulting services for CMOs on AI strategy for marketing teams through our Teamless offering.
TOOL UPGRADE
Three tools for checking AI visibility claims yourself:
BrandWell (Content at Scale) — best for B2B content teams scaling organic content with intent data and a 12-factor SEO score built in. Try BrandWell here →
Nouswise — best for research where every answer has to be checkable, because it answers from your own files and links and cites them. Try Nouswise here →
Anara — best for marketers who want to ask questions of warehouse data in plain English, without SQL or an analyst. Try Anara here →
→ We have 300+ curated, filtered, and categorized AI tools in our directory. Browse to find the best ones that help your marketing team.
LEARN THIS
Before you buy anybody's AI benchmark: The 2025 State of AI in Marketing. Ten chapters and five predictions on where marketing teams actually are with AI, not where vendors say they are.
→ We publish workshops, courses, guides, manuals, and more on AI marketing for free. Browse our learning resources for free.
BRAND SLOP
Futurism found TikTok accounts posting AI selfies of crying women beside fake texts from relatives, each carousel ending on a notification from an app. "Ashley Miller" was selling Zenit, a wellness app, and one of her posts drew more than 240,000 likes. Many commenters believed her.
None of it was labeled as AI or as an ad, Futurism reports. Zenit's developer, Pedro Pérez, owned up: "The visuals were AI-generated and the posts promoted my app, so they should have been labeled as such." He deleted the account.
Prospinity, which makes the AI companion Daimon, blamed affiliates "publishing content independently through accounts we didn’t directly manage."
If an affiliate invented a customer for you this week, would your contract even say it wasn't allowed?
Reach marketing leaders through us
You’re reading this, and so are nearly 20,000 of your peers. Some of them are your next customers.
If you sell to marketing teams, a single placement is €500 and a full week is €2,000.
If you want to place an ad, simply reply to this email or check our media kit.