The prevailing wisdom on AI search ads has been that they’re still a curiosity: a handful of sponsored links dotted through chat answers, more novelty than a genuine ad business. Test the format, watch engagement, worry about scale later.
This week that story flipped. New research from SE Ranking, analysing over 50,000 commercial ChatGPT prompts, found sponsored placements now appear on 25.94% of them, closing in fast on Google AI Mode’s 29.45%. ChatGPT isn’t dabbling in advertising any more. It is running an ad business at a scale that rivals the search giant it was supposed to be disrupting.
The pivot line: the same research shows roughly one in seven of those ads (14.35%) don’t actually match what the user asked, and in categories like relationships and politics, more than half miss the mark entirely. Scale arrived before relevance did, and marketers now have to decide how much budget to commit to a channel that’s still working out who its ads are for. Everywhere else this week, the platforms kept building anyway: OpenAI added shoppable carousels to ChatGPT, Microsoft and Google both shipped new AI insight panels for advertisers, and Common Crawl published a manual for checking whether your site is even visible to the crawlers feeding all of this in the first place.
Not every brand chased the algorithm this week, though. Foot Locker and Jif both leaned the other way, betting on storytelling and emotional connection over another round of AI optimisation.
Here is what mattered, why it mattered, and what to do with it on Monday.
Board Analysis, The Digital Maze’s weekly marketing briefing, goes deeper on all of this every Friday. Subscribe here.
Paid Media & Search: ChatGPT’s Ad Business Just Caught Up To Google. Its Relevance Problem Hasn’t.
Search Engine Land’s coverage of SE Ranking’s new study is this week’s single biggest story, because it’s the first hard data anyone has published on ChatGPT ads at scale. Across more than 50,000 commercial prompts, sponsored placements appeared on “roughly one in four commercial prompts”, landing at 25.94% against Google AI Mode’s 29.45%. For a channel that only started testing ads properly this year, that’s a startling amount of ground closed in a short time.
The relevance data is where the story gets uncomfortable. “Around 14.35% of ChatGPT ads were effectively unrelated to the prompt”, and that average hides enormous variance by category: “Just 2.6% of ads in Pets were classified as mismatched, compared with more than half in Relationships and News & Politics.” Perhaps more telling for anyone weighing whether paid placement buys credibility as well as visibility, “only 3.63% of advertisers were also cited as a source in the answer above their ad.” Buying the ad slot and earning the citation are, for now, two almost entirely separate games. For any UK marketing director sizing up ChatGPT as a media channel, the numbers say the audience and the ad inventory are both real. They also say the targeting logic is still catching up, so early budget needs monitoring at a category level, not just a platform level.
Why this matters:
- ChatGPT ads now appear on 25.94% of commercial prompts, nearly matching Google AI Mode’s 29.45%.
- 14.35% of ChatGPT ads were unrelated to the surrounding conversation, rising above 50% in categories like Relationships and News & Politics.
- Only 3.63% of advertisers who bought a placement also appeared as a cited source in the answer above it.
Read more: Study: ChatGPT ads appear on 26% of commercial prompts, Search Engine Land, 10 August 2026
Also worth your time this week:
Search & Intelligence: Half Of Marketers Are Now Funding AI Visibility. Most Still Can’t Measure It.
Digiday’s data-led roundup on AI search infrastructure is the story that best captures where budget is actually moving this week. The numbers show real commitment: “82% of marketers have allocated at least some budget, and 43% are spending more than 20%” of their search or content budget on AI visibility specifically. That’s not experimental spend any more, that’s a meaningful reallocation of existing budget toward a channel most teams have run for less than a year.
The gap is in measurement, not appetite. “67% of respondents say their brands appear less frequently in AI answers than desired”, and worse, “71% aren’t tracking share of voice against competitors” at all. The piece also flags real structural differences worth knowing before building a single “AI SEO” strategy: ChatGPT averages 15.4 sources per answer, Google AI Mode 11.4, and Gemini just 3.3, meaning the odds of being cited vary hugely by which assistant a customer happens to be using. For UK marketing leaders already spending against AI visibility, the practical gap to close isn’t budget, it’s the tracking that tells you whether that budget is working.
Why this matters:
- 82% of marketers have allocated some budget to AI visibility, with 43% spending over 20%.
- 67% say their brand appears less often in AI answers than they want, and 71% don’t track share of voice at all.
- ChatGPT cites 15.4 sources per answer on average, versus Google AI Mode’s 11.4 and Gemini’s 3.3.
Read more: By the numbers: How marketers are building the infrastructure for AI search, Digiday, 7 August 2026
Also worth your time this week:
Web Development & Design: Your Site Might Be Invisible To The AI Crawlers That Matter Most
Search Engine Journal’s piece on Common Crawl is this week’s most quietly important technical story, because it addresses infrastructure most marketing teams have never checked. Common Crawl, the non-profit web archive that underpins most LLM training data, published a manual for auditing whether a site’s robots.txt and crawl settings actually let its bot in. The scale of Common Crawl’s influence is easy to underestimate: “When Mozilla audited the LLMs released between 2019 and 2023, 64% had trained on Common Crawl data in some form, and GPT-3 was roughly 60% filtered Common Crawl by training weight.” If your site blocks that crawler, you’re not just missing one bot, you’re potentially missing the training data of a majority of major language models.
The piece’s most striking example is a household name getting it wrong: the BBC “blocks 13 of 14 tracked AI crawlers via robots.txt”, making it effectively invisible in Common Crawl’s archive despite being one of the most-cited news sources on the open web. A separate Search Engine Journal piece this week, built around a satirical fake “cats.txt” standards file, reached a similar conclusion from another angle: being crawled, indexed, and even mentioned by an LLM doesn’t prove a technical signal like llms.txt actually works, because a fabricated file that “exists nowhere but the file” passed all four common proof points. Read together, the message for technical and marketing teams is the same: check your actual crawler access directly rather than assuming a popular new standard is doing the job.
Why this matters:
- 64% of major LLMs trained partly on Common Crawl data, per Mozilla’s audit.
- GPT-3 was roughly 60% Common Crawl data by training weight.
- The BBC blocks 13 of 14 tracked AI crawlers via robots.txt, right now.
Read more: Common Crawl Published A Manual For Being Visible To AI, I Automated It, Search Engine Journal, 10 August 2026
Also worth your time this week:
Brand & Connection: When The Algorithm Gets Noisy, Some Brands Just Lean Into Feeling
Marketing Dive’s coverage of Foot Locker’s new brand platform is the clearest counter-signal to this week’s AI news, and it lands at an interesting moment for the retailer. Having returned to comparable-sales growth after its 2025 acquisition by Dick’s Sporting Goods, Foot Locker launched “It Always Will Be Foot Locker”, anchored by a two-minute hero film tracing sneaker culture back to the brand’s 1974 founding, running across broadcast, social, digital, OOH and in-store, and leading into a 36-episode documentary series made with Kevin Hart’s Hartbeat. CMO Brett O’Brien was explicit about what the campaign is actually solving: “We don’t have an issue with brand awareness… It’s really about, what does Foot Locker mean.”
Digiday’s reporting on beauty brands this week makes the same underlying point from a different category. Brands including Innisfree, Rare Beauty and e.l.f. Cosmetics are increasingly building comedic skits, mockumentaries and “Get Ready With Me” formats rather than straight product advertising, with Innisfree’s Jamie Lee arguing plainly that “with Gen Z, in particular, traditional product-focused advertising is really not enough anymore. Entertainment is 100% the way to capture an audience.” Jif’s rebrand, turning its own logo into a set of verbs to escape being pigeonholed at “only 4% of total snacking occasions”, rounds out a week where three very different brands all reached for the same lever: build feeling and relevance through story, not another optimisation pass.
Why this matters:
- Foot Locker’s CMO says the brand isn’t short on awareness, only on clear meaning.
- The campaign spans broadcast, social, OOH and in-store, plus a 36-episode documentary series.
- Jif turns its logo into verbs to escape a 4% share of snacking occasions.
Read more: Foot Locker reinforces role in sneaker culture with new brand platform, Marketing Dive, 7 August 2026
Also worth your time this week:
AI & Automation: Every Major Ad Platform Shipped An AI Update This Week
Digiday’s report on OpenAI’s new carousel ad format ties this week’s platform news together neatly. ChatGPT can now show multiple products from one retailer’s feed in a single ad unit, extending the single-product feed format OpenAI introduced three months earlier, with OpenAI’s own system, not the advertiser, deciding which format to serve. Agency sources quoted in the piece are already thinking ahead to Q4, with one Adthena executive noting “there’ll be even more pressure in Q4 because advertisers will be planning [budgets] for 2027” against reported targets of $2.5 billion in ad revenue this year scaling past $100 billion by 2030.
OpenAI wasn’t alone. Microsoft’s first monthly product newsletter this week bundled a new “Topic Insights” feature inside Clarity’s AI visibility reporting, two new Performance Max experiment types, and an expanded Ad Preview Hub, backed by Microsoft’s own claim of “an average 8% increase in incremental conversions from Performance Max campaigns.” Google, meanwhile, added AI Overview summaries to the top of the Analytics homepage and prompt-driven insight cards to Google Ads, with Google framing it as giving marketers “the context you need to confidently adjust your strategy, ensuring you capture new demand the moment it appears.” Three platforms, three separate AI ad updates, in the same seven days. The direction of travel for any UK marketing team running paid media is unambiguous, even if the winning playbook inside each platform is still being written in real time.
Why this matters:
- OpenAI’s system decides single or carousel ad format, not the advertiser.
- Microsoft’s new PMax experiments show an average 8% lift in incremental conversions.
- OpenAI reportedly targets $2.5bn in ad revenue in 2026, scaling past $100bn by 2030.
Read more: OpenAI brings product carousels to ChatGPT ads, Digiday, 6 August 2026
Also worth your time this week:
Case Study of the Week: 600% More AI Engine Traffic For A Supplement Brand Nobody Was Finding Online
This week’s theme, that AI visibility infrastructure is now something every brand needs to actively build rather than assume, is exactly what The Digital Maze delivered for ParkAcre, a specialist supplement manufacturer that came to us after an earlier web development project needing a data-driven SEO strategy from close to zero organic visibility. We built out long-tail and intent-based content, dedicated market-segment pages, e-commerce catalogue work, technical and internal-linking improvements, and optimised content specifically for AI-driven search engines rather than treating it as an afterthought.
- +85% increase in SEO-generated leads, year on year
- +83% growth in organic sessions, year on year
- +316% rise in first-page keyword rankings, from 118 to 382 keywords
- +600% surge in AI Engine traffic, with a +666% increase in AI Engine-associated leads
Read the full ParkAcre case study.
Bringing It Together
Every story this week is really about the same tension: AI search and AI ads are scaling faster than anyone’s ability to measure or control what’s actually happening inside them. ChatGPT’s ad inventory grew to rival Google AI Mode’s before its relevance targeting caught up. Half of marketers are funding AI visibility without a reliable way to track it. A household-name publisher is invisible in the archive most LLMs train on, without realising it. And every major ad platform shipped a new AI tool this week, each one asking marketing teams to trust a system that’s still learning as it goes.
The brands that stood apart this week didn’t fight that uncertainty with more optimisation. Foot Locker, the beauty brands profiled by Digiday, and Jif all leaned into story and feeling instead, on the logic that emotional connection is a more durable asset than any single platform’s current algorithm. None of this week’s lessons require abandoning AI tools or AI advertising. They require checking, deliberately and often, whether the systems making decisions on your behalf, about relevance, about visibility, about crawler access, are actually doing what you assume.
The practical takeaway is to treat AI visibility the same way you’d treat any other line item you can’t yet fully measure: worth investing in, but not worth trusting blindly.
Three Takeaways for the Week Ahead
- Audit whether your AI ad placements actually match user intent. 14.35% of ChatGPT ads were flagged as unrelated to the prompt this week, with mismatch rates above 50% in some categories.
- Stop over-segmenting your PPC accounts. Segments need 30 conversions in 30 days to give bidding algorithms a usable pattern, or you’re just fragmenting your own data.
- Check your site’s crawlability for AI, not just Google. The BBC blocks 13 of 14 tracked AI crawlers via robots.txt, meaning it’s effectively invisible in the archive most LLMs train on.
Want help putting any of this into action?
We work with ambitious brands across SEO, PPC, web development, and content strategy. If this week’s news has raised questions about your AI visibility, your PPC account structure, your site’s crawlability, or how well your brand connects beyond the algorithm, we would like to talk.
Get in touch with The Digital Maze.
Board Analysis is The Digital Maze’s weekly marketing briefing, published every Friday. It covers search, paid media, web development, brand, and AI for marketing managers, directors, and leaders who need to stay ahead without wading through every trade publication themselves.
