The Brief: AI Trusts Earned Media, A $20 Tool Wants Your Agency's Job, Meta Wants In On the Cloud Business, and Challenger Brands Are Winning With Less
In today’s brief: Four stories this week that matter if you're thinking seriously about B2B marketing in the AI era.
Today's stories:
1. AI visibility depends on who writes about your brand, not just where you rank
2. Can a $20 AI tool actually replace your marketing agency
3. Meta is reportedly building a cloud business to compete with AWS and Google Cloud
4. How challenger brands like Emerald Nuts win with fewer, bigger marketing bets
Rankings Get You Traffic, Earned Media Gets You Cited
Quick gut check for you. BrightEdge tracked a year of data and found that only about 17% of sources cited in AI Overviews also rank in Google's organic top 10 for that same query. Moz ran its own analysis of 40,000 AI Mode queries and came back with an even starker number, with 88% of citations pulled from pages that never cracked the top 10 at all. So yes, you can hold the number two spot on Google and still be invisible to anyone asking an AI for a recommendation.
So what actually gets you cited then? Earned media, and it's not particularly close. Muck Rack's Generative Pulse team dug through more than 25 million links cited across ChatGPT, Claude, and Gemini and found earned media accounts for 84% of all AI citations, while paid and ad content barely register at 0.3%. Journalism alone makes up 27% of cited sources, a share that's held steady since July 2025. Named, credentialed authors help too.
This is one of the lanes I work in for brands and executives who want to land in that 84%. I offer is executive and personal branding services, which includes pitching expert commentary on a brand's behalf to trade publications and industry outlets so their name becomes one of the sources AI systems are actually citing. If nobody's quoting a named expert from your brand anywhere, you're simply not part of the conversation these systems are having about your category.
Read more: https://martech.org/ai-visibility-depends-on-who-writes-about-your-brand/
Can a $20 AI Tool Actually Replace Your Agency?
You've probably scrolled past the LinkedIn post already. Someone cancelled a $5,000-a-month agency, spent $20 on an AI tool instead, and now claims a robot handles their SEO while they sleep, complete with a screenshot of a suspiciously tidy score out of 100. Joanna Lambadjieva, who writes The AI Search Optimization Newsletter, decided to actually dig into whether that's true. Her answer lands somewhere in the middle of that argument, and it's worth hearing out.
According to Joanna, the tools are genuinely good at the boring, mechanical stuff. Site speed problems, pages accidentally hidden from Google, technical issues quietly costing you sales, an AI tool will catch all of that faster than any human and never get bored doing it. She also points out they're useful for tracking whether your brand shows up when someone asks ChatGPT for a recommendation, which used to mean manually testing dozens of prompts a week and now just runs quietly in the background.
But maybe Joanna's main argument points to judgment. These tools tend to flag twenty problems as equally urgent because they can't tell what actually matters, and most only glance at a handful of pages before scoring your entire site, missing exactly where the real problems live if you're running hundreds of similar product pages.
They're also nearly silent on backlinks and reputation, which happen to be some of the biggest factors in whether you rank at all. Her take, and I'd agree, is that the tools are a legitimately good deal for the mechanical work, but the judgment calls that actually grow a business are still a human job.
Read more: https://aisearchoptimization.beehiiv.com/p/can-a-20-ai-tool-replace-your-agency
Meta Wants In on the Cloud Business Too
Meta wants a slice of the cloud business as much as I want the Bills to win the Super Bowl (a lot). According to a Bloomberg report, the company is quietly building an infrastructure business that would sell AI computing power and model access to outside customers, which would put it in direct competition with AWS, Google Cloud, and Microsoft Azure.
One version of the plan looks like AWS Bedrock, letting developers tap into AI models hosted on Meta's own servers. The other looks more like neocloud players such as CoreWeave, just selling raw computing capacity outright.
Zuckerberg has actually been hinting at this for a while. On Meta's May earnings call, he said companies ask to buy Meta's spare compute almost every week, at a premium, and Meta just hasn't said yes yet because it's still using that capacity itself. If Meta ever decides it's overbuilt, monetizing the leftovers becomes an easy call. It fits the pattern we're seeing across Meta's AI bets lately, testing every possible way to make money off this buildout, including a rumored $199.99-a-month premium tier for its Hatch AI agent that would put it in the same weight class as Anthropic and OpenAI at the high end.
Read more: https://www.pymnts.com/technology/2026/meta-plans-cloud-business-to-take-on-big-tech-rivals/
Challenger Brands Are Winning by Doing Less, on Purpose
Dan Eisenberg, CMO at Chicago agency Blue Chip, has a philosophy for challenger brands competing with corporate giants and ten times the budget. He calls it "bigger, fewer, better," which basically means making disciplined, narrow choices about which audiences, markets, and retailers actually matter, then going all in there instead of spreading thin everywhere.
Every dollar also has to pull double duty, building the brand and driving the sale at the same time, because challenger brands typically can't afford to run separate branding and performance strategies.
For example, Bob's Red Mill went all data-driven, using a tool that algorithmically scored which markets were actually worth the investment, then poured spend there instead of chasing broad national reach. The brand posted its first-ever million-dollar week at Walmart in Q4 2025, part of five straight years of household penetration growth against giants like Quaker Oats and King Arthur Baking.
That's it for today.
Follow me on LinkedIn and BlueSky, and X
FAQs
Does ranking well in Google guarantee AI visibility?
No. BrightEdge found only about 17% of sources cited in AI Overviews also rank in Google's organic top 10, and Moz found 88% of AI Mode citations came from pages outside the top 10 entirely. Traditional SEO rankings and AI citations are increasingly separate systems.
What content do AI systems trust most when generating answers?
Earned media, by a wide margin. Muck Rack's Generative Pulse analysis of 25 million cited links found earned media accounts for 84% of all AI citations, compared to just 0.3% for paid and advertorial content. Journalism alone makes up 27% of cited sources.
Can AI tools replace an SEO or marketing agency?
Partially. AI tools are strong at technical audits and monitoring AI search visibility, but they struggle with prioritization, judging actual content quality, and assessing backlinks or reputation, which remain some of the biggest factors in search performance.
Is Meta building a cloud computing business?
Yes, according to a Bloomberg report. Meta is reportedly developing a cloud infrastructure business under its Meta Compute initiative that would sell AI computing power and model access to outside companies, competing directly with AWS, Google Cloud, and Microsoft Azure.