AI search has been simple: get cited, stay visible, watch your mentions climb. Solve the visibility problem and the rest follows.
This week the problem turned out to sit one step earlier than visibility. New research covering 71 verified businesses found the average company doesn’t just struggle to rank inside AI answers, it actively “leaks” 84% of its own identity before an AI system ever gets the chance to describe it properly. Seventeen percent of the sample had no AI-retrievable presence whatsoever, and only 22 of the 71 businesses had identifiable leadership visible to AI at all.
The thesis for the week: control over how a brand gets represented, whether by an AI model, an ad platform, or its own campaign spend, is shifting into places most marketing teams haven’t thought to audit yet. Google admitted its Performance Max campaigns had become enough of a black box that advertisers were openly asking for an off switch, and quietly built one. Anthropic let Google index thousands of private Claude conversations. And M&S reminded the industry that five years as Britain’s most trusted brand came from consistency, not campaign spend.
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.
Search & Intelligence: Most Brands Have No Idea What AI Thinks They Are
Search Engine Land’s new six-dimension framework for auditing your “AI entity footprint” arrives at a genuinely useful moment. The framework, built around Identity, Differentiation, Evidence, Consistency, Relationships and Specialization, treats the question of how AI systems understand a business as a distinct discipline from traditional page-level SEO. Its core recommendation is refreshingly simple: literally ask ChatGPT, Gemini and Perplexity to describe your business, then score the response against each dimension, since as the piece puts it, “every webpage, review, business profile, article, podcast appearance, speaking engagement, and third-party mention contributes information that may shape how AI systems understand the business.”
That framework lands in the same week as two cautionary tales about identity control going wrong in opposite directions. Private Claude AI conversations became publicly searchable on Google and Bing after Anthropic combined a robots.txt block with a noindex tag on the same pages, a configuration that fails silently: as SEO consultant Glenn Gabe explained in Search Engine Land’s coverage, “if you block via robots.txt AND noindex the page, Google and Bing cannot see the noindex tag since they can’t crawl the page and see the tag in the HTML.” Meanwhile, a US federal court dismissed two of Google’s own DMCA claims against scraping tool SerpApi, ruling Google hadn’t shown its anti-scraping system was implemented “with the authority of the copyright owner,” a live reminder that the infrastructure deciding what gets crawled and indexed is still being argued over in court.
Why this matters:
- Score your own brand this week by asking ChatGPT, Gemini and Perplexity to describe it against the six-dimension framework.
- The “Evidence” dimension rewards third-party validation like case studies and reviews over more on-site copy.
- Any brand offering shareable AI-generated links or reports should check for the exact robots.txt-plus-noindex failure that exposed Claude chats.
Read more: How to audit your AI entity footprint, Search Engine Land, 27 July 2026
Also worth your time this week:
Paid Media & Performance: Google’s Black Box Just Got A Crack In It
Digiday’s reporting on Google quietly piloting new exclusion controls for Performance Max is the biggest concession to advertiser demands since PMax launched. Media buyers can now, in early access, exclude Google Display Network and third-party Search Partner inventory from PMax campaigns. Per PMG’s Kaitlin McGrew, better reporting and control had already “led to a lift in PMax spending,” with clients shifting budget from Shopping into PMax for “a 10% average increase in PMax investments.” Duncan Channon’s Kyle Rovinski summed up why the old approach frustrated buyers: “you could not trust it.” The timing isn’t a coincidence either. Alphabet’s Q2 earnings call, covered the same week by Search Engine Land and Marketing Dive, showed search revenue growth cooling to 17% from 19% a year earlier, even as total ad revenue hit $81.6 billion, up 14.5% year on year, boosted by World Cup viewership and Gemini-driven relevance gains.
The rest of the pillar reads as Google managing a genuine platform transition rather than resisting one. Search Engine Land’s coverage of the Dynamic Search Ads sunset confirms Google has pushed the forced migration to AI Max back to February 2027, giving advertisers until January 2027 to keep running standalone DSA campaigns if they choose. As the piece notes, “in standalone DSA, the website is the foundational targeting signal. In AI Max, the website is just one of many targeting signals, including existing keywords, ad copy, budget, and real-time user intent signals.” Meanwhile Digiday reported that OpenAI is actively staffing up a full advertising sales operation, with internal targets reportedly calling for “$2.5 billion in ad revenue for 2026, and $100 billion by 2030,” though eMarketer’s Nate Elliott forecasts OpenAI will “fall roughly 90% short of that target.” Read together, the message for UK marketers is that the era of one dominant, opaque automation model is ending, and multiple platforms are now actively competing for a more transparent, better-controlled slice of ad budgets.
Why this matters:
- Ask your Google account rep about early access to the new PMax exclusion controls, since agencies report a queue for access.
- The DSA migration deadline has moved to February 2027, giving accounts extra time to plan rather than being force-migrated this September.
- ChatGPT ad inventory remains a test-budget channel, not one to reallocate core PPC spend into yet, per eMarketer’s own scepticism of OpenAI’s targets.
Read more: Google quietly gives ground on PMax controls, Digiday, 24 July 2026
Also worth your time this week:
Web Development & Design: The 5-Second Rendering Rule Was Never Real
A new test from Dave Smart at Tame the Bots, covered by Search Engine Journal, disproves a technical SEO assumption that has shaped development decisions for years: that Google’s Web Rendering Service cuts off rendering after five seconds. Smart found Google’s crawler uses a “virtual clock” that pauses while waiting on network requests, meaning it successfully captured content from delayed server-side API calls that took “between six and twelve seconds” combined, well past the supposed cutoff. Google’s own Martin Splitt clarified that the original five-second figure referred to the median time a page waits in the rendering queue, not an actual rendering time limit, “a distinction that was lost in translation across the SEO industry.”
For any UK marketing team running a JavaScript-heavy front end, whether React, Vue or a headless CMS setup, this is a genuinely practical correction. Development teams have spent real budget rushing to “beat a five-second clock” that doesn’t functionally exist in the way the SEO industry assumed. It doesn’t mean slow sites are fine (Core Web Vitals still matter for user experience and conversion), but it does mean the rendering-cutoff myth shouldn’t be the reason a headless architecture or delayed API call gets vetoed by an SEO team acting on outdated advice.
Why this matters:
- Google’s crawler pauses its “virtual clock” while waiting on API responses, rather than abandoning rendering outright.
- Content arriving 6 to 12 seconds after initial page load was still successfully captured and indexed in the test.
- JavaScript-heavy React, Vue and Next.js stacks are not automatically penalised by a hard rendering cutoff that doesn’t exist.
Read more: Google SEO Test Shows What Happens In 5-Second Rendering Window, Search Engine Journal, 28 July 2026
Also worth your time this week:
Only one article fully cleared this week’s verification bar for the web development and design pillar. Coverage across the whitelist skewed heavily toward AI search and ad-platform news this week rather than technical foundations or UX, so no padding candidates are included here.
Brand & Connection: The Ordinary Just Proved Mocking Your Own Industry Builds Trust
The Drum’s account of The Ordinary and creative agency Uncommon’s Cannes Grand Prix-winning “The Periodic Fable” campaign is this week’s best case study in earned trust. Rather than simply repeating its affordability message, The Ordinary replaced periodic table elements with skincare marketing buzzwords like “medical grade” to publicly mock the industry’s own gimmicks. The campaign reached “85 million people overall,” and, per the article, “58% of US consumers said the campaign has made them more likely to see The Ordinary as the most trustworthy beauty brand,” alongside a “10% uplift in purchase intent.” As co-founder Nils Leonard frames the method, the trick is to “dance on tension”: find the industry buzzword everyone quietly rolls their eyes at, then interrogate it publicly.
Two further Drum interviews this week reinforce the same underlying idea from different angles. M&S marketing director Sharry Cramond told The Drum that “campaigns are less than 10% of marketing,” crediting sustained brand-building, not individual ad bursts, for the fact that “M&S is now the most loved and trusted brand in Britain, and we’ve had that accolade for the last five years.” Meanwhile EY’s outgoing global chief brand and marketing officer John Rudaizky argued B2B buyers are more emotionally invested than marketers assume, not less: “picking up a detergent brand is not as big a decision… something that’s going to affect your future, your family, your mortgage, that’s a different scale of emotional decision-making.” Under his 13-year tenure, EY’s brand climbed “from 75th to a top-10 AAA+ brand in Brand Finance’s rankings.” Read together, all three stories say the same thing: trust compounds through consistency and creative courage, not through a single clever campaign or a rational features list.
Why this matters:
- 58% of US consumers found The Ordinary more trustworthy after a campaign that mocked category jargon rather than just claiming affordability.
- M&S has held the “most loved and trusted brand in Britain” title for five straight years while treating campaigns as under 10% of its marketing effort.
- EY’s brand climbed from 75th to a top-10 AAA+ ranking over 13 years by leaning into emotional B2B messaging, not against it.
Read more: “Dancing on tension”: How The Ordinary and Uncommon are pushing category norms, The Drum, 24 July 2026
Also worth your time this week:
AI & Automation: 500,000 Advertisers Are Already Running On AI Max
Alphabet’s Q2 earnings call was the single biggest news event of the week, and Search Engine Land’s coverage of the AI Max numbers is the most consequential detail buried inside it. Google confirmed AI Max has exited beta and been adopted by over 500,000 advertisers, with those running AI Max or Performance Max seeing an average 15% lift in conversions or conversion value at a similar return on ad spend. Google framed the shift as unlocking “billions” of previously unmonetised searches, by matching ads to complex, conversational queries that traditional keyword targeting structures were never built to handle.
Two supporting stories fill out the picture. Marketing Dive’s earnings coverage confirmed Google’s total ad revenue hit $81.6 billion for the quarter, up 14.5% year on year, with YouTube ad revenue climbing 13% partly on the back of 1.7 billion unique World Cup viewers globally, while Alphabet’s 2026 capital expenditure guidance of $195 to $205 billion signals continued heavy AI infrastructure investment. Separately, Search Engine Land reported AI Max features are now rolling into Standard Shopping campaigns too, a format advertisers had deliberately kept separate from Performance Max to retain manual control. Retailers who stayed on Standard Shopping specifically to avoid PMax’s automation now need to review their settings, since the rollout brings automatic ad copy generation and Final URL Expansion with it by default. For UK marketing teams, the throughline is straightforward: AI-driven automation is no longer an emerging feature of Google Ads, it is the majority behaviour, at scale, backed by real earnings numbers.
Why this matters:
- Over 500,000 advertisers have already adopted AI Max, with a 15% average conversion lift reported at similar spend levels.
- Retailers on Standard Shopping specifically to avoid PMax automation should review settings now that AI Max features are rolling in there too.
- Alphabet’s $195-205bn 2026 capex guidance signals AI-driven ad tooling will keep evolving fast enough to require ongoing team retraining.
Read more: Google says AI Max unlocks billions of new monetizable searches, Search Engine Land, 23 July 2026
Also worth your time this week:
Case Study of the Week: 168% More Conversions Without A Bigger Budget
This week’s theme, that trust and visibility come from consistent fundamentals rather than a single lever, is exactly what played out for TwentyCi, which came to The Digital Maze wanting more than vanity traffic. We combined an SEO-led content and keyword strategy to grow organic reach, targeted PPC and remarketing campaigns to capture qualified leads, and email marketing that reinforced TwentyCi’s industry expertise with its existing contacts.
- +168.86% increase in goal completions
- +144.61% increase in goal conversion rate (2022 vs 2021)
Read the full TwentyCi case study.
Bringing It Together
Every story this week is really about the same shift: the entities that used to control how a brand gets seen, whether that’s an AI model, an ad platform, or a single campaign, are being forced to give some of that control back, or are quietly failing at holding onto it. AI systems can’t verify most businesses because those businesses haven’t fed them the evidence. Google’s Performance Max black box cracked open because advertisers stopped tolerating it. Anthropic’s own sharing settings failed because a robots.txt rule and a noindex tag were never tested together.
The optimistic read, hiding in plain sight, is that every one of these gaps is fixable with unglamorous, achievable work: run the AI entity audit, ask for early PMax access, check your own sharing settings for the same silent failure that hit Anthropic. None of it requires a bigger budget. It requires treating “how are we represented” as an operational question, not just a creative one.
The practical takeaway isn’t to panic about AI or automation. It’s to audit the plumbing behind your brand’s identity with the same rigour you’d apply to a paid media account.
Three Takeaways for the Week Ahead
- Audit your own AI entity footprint this week. Ask ChatGPT, Gemini and Perplexity to describe your business and score the answer against the six-dimension framework, since 84% of businesses studied are leaking their identity somewhere.
- Ask your Google rep about early access to PMax’s new exclusion controls. Advertisers already using them are cutting junk inventory from Performance Max without giving up automation entirely.
- Stop treating campaigns as your only trust-building mechanism. M&S has been Britain’s most trusted brand for five years running while spending less than 10% of its marketing effort on campaigns. Consistency, not bursts, is what compounds.
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 paid media budget, or how consistently your brand is building trust, we would like to talk.
Get in touch with The Digital Maze.
Board Analysis is The Digital Maze’s weekly marketing briefing, published every Wednesday. 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.
