There is a version of this week where you feel quietly smug about your Google rankings. You have done the work. You are on page one. The traffic is coming in. Job done.
Then research lands showing that the median enterprise B2B brand appears in just 3% of AI Overviews, despite ranking well for thousands of keywords. The buyer who asks ChatGPT, Perplexity, or Google AI Mode about your product category has a 97% chance of receiving an answer that does not mention you at all.
That is the central tension this week. Not AI versus human creativity, though Cannes Lions delivered a verdict on that too. Not new Google ad formats, though there were several worth paying attention to. The real theme is this: the signals that made your brand visible in traditional search have almost no bearing on whether AI systems include you in their answers.
Twenty thousand marketers flew to Cannes to debate this and related questions. Google released a spam update while everyone was distracted. Amazon unveiled an ad format that completes the transaction inside a voice conversation. And a piece of research quietly confirmed that 57.2% of all web traffic is now from automated bots, yet 95.9% of websites have accessibility failures that make them effectively unreadable to those bots.
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: The AI Visibility Gap Just Became Measurable
The most consequential piece of research published this week came from Search Engine Land, reporting on a study of 828 enterprise B2B companies across 45 million search queries. The headline number: the median brand appears in just 3% of AI Overviews, despite AI Overviews appearing in 50% of searches where those brands rank organically. The top quartile of brands reaches only 4.5%. For professional services and distribution companies, the figure drops to 2.1%.
To put that in concrete terms: a business ranking for 9,700 keywords is cited in AI responses for roughly 290 of them. The other 9,410 searches where the brand ranks traditionally produce AI answers that the buyer sees first, and that brand is absent from them. The measurement gap compounds the problem. Fewer than 26% of marketers currently track AI visibility, meaning most businesses are optimising for a metric that no longer captures the full buyer research journey.
The study also found that 4.6% of enterprise B2B companies receive zero AI citations across all their relevant keywords. Not 3%. Zero. The implication is uncomfortable: years of SEO investment have built a Google presence that AI systems have largely chosen to ignore.
What separates the brands that do get cited? Deep topical authority rather than broad keyword coverage. Structured explanations. Direct answers to specific questions. Cybersecurity brands lead all industries at a 4.2% citation rate, a sector that has historically published detailed, technically precise content. That pattern is not coincidental.
Meanwhile, Google confirmed the June 2026 spam update completed on 26 June, described as "a bit bigger than March 2026." Sites relying on manipulative link-building tactics will have seen ranking drops from 24 June. The practical note: sites previously benefiting from spammy backlinks will not regain their lost ranking value even after removing those links. Recovery requires building genuine authority, the same thing the AI citation research rewards.
Why this matters:
- Ranking well on Google and appearing in AI Overviews are now two separate problems requiring two separate strategies. The data confirms there is almost no correlation between the two.
- The measurement gap is a strategic risk. If your team is not tracking AI citation rates alongside traditional rank positions, you are flying blind on a growing share of your buyers' research behaviour.
- Content depth and structure matter more than volume. Publishing 20 blog posts on tangentially related keywords will not build AI citation presence. One definitive, well-structured piece on a core topic will.
Read more: B2B brands rank in Google but appear in just 3% of AI Overviews , Search Engine Land, 24 June 2026
Also worth your time this week:
Paid Media & Performance: Google Continues Handing More Over to AI
Three significant updates landed for Google Ads practitioners this week, each pointing in the same direction: Google is systematically reducing the levers advertisers can pull manually while expanding what its AI systems do automatically.
The largest update went to Demand Gen campaigns. Google announced that Gemini will now provide automated recommendations on how to optimise creative assets for YouTube before campaigns go live, alongside automatic video resizing that transforms content between vertical, square, and landscape formats without additional production. The update also adds web-to-app conversion measurement, allowing advertisers to track when web campaigns lead users to install an app. The headline statistic from Google's own data: 72% of incremental conversions on YouTube come from new customers, per research from Measured. If Demand Gen is in your mix, this measurement update directly affects how you report acquisition performance to a board or budget holder.
Google also updated the Ads API to version 24.2, adding SyntheticContentInfo and SyntheticContentAttestation fields that allow advertisers to identify and label AI-generated creative programmatically. The practical reason to care is timing: the EU AI Act's requirements for disclosure of AI-generated ad content take effect in August 2026. Any advertiser or agency building campaigns at scale needs to understand these fields before that deadline.
The third update is the one that will most directly change day-to-day campaign management. Google has rolled out Maximize Conversion Value bidding to Standard Shopping campaigns without requiring a Target ROAS setting. Previously, accessing value-based bidding required either setting a hard ROAS target or switching to a feed-only Performance Max campaign, a workaround many advertisers found limiting. That workaround is now unnecessary.
Why this matters:
- Gemini-powered creative recommendations before launch reduce the trial-and-error cost of testing YouTube creative, particularly valuable for teams running Demand Gen with limited creative budgets.
- The EU AI Act compliance deadline of August 2026 is closer than most teams have planned for. If you are using AI-generated ad creative at scale, the API v24.2 updates are not optional reading.
- The Standard Shopping bidding update removes a key reason advertisers moved reluctantly to Performance Max. Teams who built feed-only PMax campaigns purely for Maximize Conversion Value bidding can now simplify their account structures.
Read more: Google gives Demand Gen new AI creative and reporting tools , Search Engine Land, 25 June 2026
Also worth your time this week:
Web Development & Design: The Machines Are Reading Your Site. They Cannot Understand It.
The most practically urgent story in this week's web development news was also the least visible. Research published via Search Engine Journal, drawing on data from May-June 2026, found that 57.2% of all HTTP requests to HTML content now come from automated bots, surpassing human traffic at 42.8%. At the same moment, 95.9% of the top million homepages have detectable WCAG accessibility failures, averaging 56 errors per page. That is a 10.1% increase in errors from 2025.
The connection is not coincidental. AI agents do not read your website the way users do. They read the accessibility tree: a semantic structural model that browsers generate from your DOM code. The same tree that powers screen readers is what Google's Gemini agent, ChatGPT's browser, and Amazon's Alexa use to understand what your page says and what it means. When 83.9% of sites have low-contrast text, 53.1% are missing alt text, and 51% are missing form labels, those agents encounter your brand as a structurally incoherent document, not a credible business.
The research found something counterintuitive about ARIA, the set of attributes designed to improve accessibility. Pages using ARIA averaged 59.1 errors compared to 42 on pages without it, suggesting that misapplied ARIA attributes are actively making machine comprehension worse. The recommendation is to use native HTML elements rather than patching broken markup with ARIA. Agents and screen readers both prefer it.
Google's John Mueller confirmed this direction explicitly on 26 June, answering a question about whether SEO principles change for AI agents. His answer: the fundamentals do not change, but blocking agentic browsers carries SEO risk comparable to blocking Googlebot. "Some details will undoubtedly evolve," he wrote, noting that not blocking agentic browsers is now a basic technical requirement in the same way allowing crawling has always been. The historical parallel he drew, to nofollow links creating unintended SEO consequences, is an instructive one.
Why this matters:
- Your website's accessibility code is now your AI readability layer. A site that AI agents cannot parse accurately is a site those agents will not cite accurately in search results or shopping interfaces.
- Misapplied ARIA worsens machine comprehension. If your development team has been adding ARIA attributes as a quick fix, it is worth auditing whether those additions are helping or creating additional noise.
- Blocking AI agents carries SEO risk. John Mueller's comments confirm that site configurations designed for human-only crawling need to be reviewed before Chrome's auto-browse feature, announced for late June launch on Android, becomes widespread.
Read more: The Accessibility Tree Is How AI Agents Read Your Site & It's Breaking , Search Engine Journal, 24 June 2026
Also worth your time this week:
Brand & Connection: Cannes Voted for Craft. The Data Agreed.
Cannes Lions 2026 finished on 26 June with a verdict that should matter to any marketing director signing off on a creative budget. About 40% of submissions used AI in some form. The technology dominated the beach activations and the press coverage. And yet, as Rebecca Sykes noted from inside the jury rooms, "There's a lot of tech presence, there's not a lot of technological 'wow.'" The work judges rewarded was not the work that scaled fastest. It was the work that was made best.
Coinbase won the Grand Prix in Film Craft with a campaign that used practical special effects and retro video game aesthetics. Dove won with unfiltered Reddit reviews, the kind of user-generated honesty that cannot be generated at scale because it requires real customers to have genuinely strong feelings. Both campaigns succeeded not despite their constraints but because of them. The creativity came from editorial judgement, not computational throughput.
The other structural shift at Cannes was the creator economy reaching critical mass. US ad spending on creators forecasts $44 billion in 2026, and the discussion shifted from whether creators should be part of the media mix to how to integrate them into the broader marketing ecosystem. "Everyone's in an arms race to demonstrate value from AI utilisation," noted Angela Tangas of Oliver, but the 500 creators who attended Cannes were drawing larger audiences than many of the keynote sessions on AI tools.
The data for challenger brands reinforced the same conclusion from a different angle. Emerald Nuts drove an 11% engagement increase among millennial shoppers and a 30% boost in new-to-brand purchases through its first national TV campaign in years, repositioning around quality using humour and a focused media buy. Bob's Red Mill recorded its first million-dollar week at Walmart in Q4 2025 through geographic expansion and a single consistent brand message. CMO Dan Eisenberg summarised the principle as "fewer, bigger, better," and the phrase captures exactly what Cannes was rewarding: disciplined focus rather than diversified volume.
Why this matters:
- AI is a production tool, not a creative strategy. The brands winning at Cannes and in market share are using AI to reduce friction while investing more, not less, in editorial judgement and craft.
- Creator marketing is now a core channel, not a supplementary one. $44 billion in US creator ad spend is not a trend. It is a structural shift in where audience attention and trust live.
- Challenger brand economics require every pound to do double duty. Brand-building and direct response are not separate budget lines. The campaigns producing returns treat them as one.
Read more: 5 Cannes Lions takeaways as marketers wrestle with an industry in flux , Marketing Dive, 26 June 2026
Also worth your time this week:
AI & Automation: Agentic Commerce Is Live. Not Coming. Live.
The most practically significant AI announcement at Cannes was not OpenAI's positioning of ChatGPT advertising or Meta's announcement about content moderation. It was Amazon's unveiling of Alexa+ Agentic Ads: a format that allows consumers to complete purchases inside a voice conversation on Echo Show devices without navigating to a website, an app, or a payment screen.
Amazon confirmed that Papa Johns and The Orchard, a Sony Music subsidiary, are already live in beta. Customers can discuss options with Alexa, receive personalised recommendations based on conversation history (including, if you order from Papa Johns regularly, remembered pizza preferences), and complete the transaction entirely within the interface. Charlotte Maines, Amazon's VP of devices content and advertising, was unambiguous at Cannes: "We have ad products that are live today. This isn't something that we're talking about or 'we're working on it.'"
The advertising industry's verdict at Cannes was sceptical but attentive. Mike Feldman of Flywheel described paid agentic ads as "the tax you have to pay for messing up on the organic side," framing the priority correctly: brands that build strong AI-native presence through structured data, conversational content, and accessible markup will need to buy less reach through paid agentic formats. The parallel with paid versus organic search is instructive, and the window to build organic AI presence before paid competition drives up costs is probably shorter than most teams are planning for.
Elsewhere in AI automation: Meta confirmed it plans to reach 90% AI automation of content moderation across Facebook and Instagram by end of 2026, up from 50% currently. And for marketing teams with a practical interest in AI tools, Search Engine Land published a guide to six content audit workflows built in Claude, covering brand voice consistency, freshness audits, AEO retrievability scoring, and topical gap analysis. Each workflow is concrete enough to implement this week.
Why this matters:
- Agentic commerce is not a pilot. It is in market. Brands selling through Amazon have a narrow window to understand how Alexa+ Agentic Ads work before the format scales and competition drives up costs.
- The organic-first principle applies to agentic AI as it does to search. Building genuine AI brand presence through structured content and accessible site architecture reduces long-term dependency on paid AI placements.
- AI content moderation at 90% on Meta's platforms changes the content moderation landscape for advertisers. Appeals processes, ad disapproval patterns, and community standards enforcement will all be handled differently.
Read more: As agentic advertising floods Cannes, Amazon touts its Alexa advantage , Marketing Dive, 25 June 2026
Also worth your time this week:
Case Study of the Week: 141% More Leads from a Joined-Up Strategy
This week's themes, AI visibility, creative quality, and the need for every marketing pound to do double duty, are exactly what drove results for KLG Rutland, an East Midlands door, window, and conservatory supplier that came to The Digital Maze in July 2022.
KLG Rutland was not struggling. They had brand recognition in their market and an existing digital presence. What they lacked was a strategy that made all the parts work together: paid search, organic search, social advertising, and local visibility all operating from the same set of goals rather than as separate channels with separate metrics.
The Digital Maze built an integrated programme combining Google Ads, Microsoft Bing Ads, Facebook Ads, and a monthly SEO programme covering content optimisation, technical improvements, and local search signals. The results measured across the full programme:
- 141% overall increase in leads from the combined strategy
- 61% increase in PPC conversion rate
- 28% increase in paid goal completions
- 19% reduction in cost per goal completion
- 130+ page one Google rankings maintained throughout the programme
The lesson from KLG Rutland is the same one challenger brand research confirmed at Cannes this week: joined-up strategies that build the brand while driving immediate conversions produce better economics than either channel operating in isolation.
Read the full KLG Rutland case study.
Bringing It Together
The thread connecting this week's news is a single uncomfortable truth: the digital presence most marketing teams have spent the last decade building was optimised for a world that is changing faster than most strategies can adapt to.
Google rankings were the measure of visibility. Now they are one measure, and they do not carry over reliably into AI search results. Website quality was measured by page speed and Core Web Vitals. Now the most important technical question is whether AI agents can parse your accessibility tree. Creative effectiveness was measured by CTR and conversion rate. Now Cannes is confirming that the most effective creative at Cannes is built on human craft, not production velocity.
None of this makes the previous decade's work irrelevant. Strong Google rankings still drive traffic. Good website performance still matters for users and crawlers. Solid direct-response creative still converts. But each of these has a new dimension attached to it, and teams that do not start building for the new dimension will find themselves competitive today and structurally disadvantaged in 12 months.
The answer is not to abandon what is working. It is to build the additional layer: AI visibility measurement, accessibility-led technical development, and creative that earns authentic response rather than just generating sufficient volume.
Three Takeaways for the Week Ahead
- Start measuring AI visibility alongside traditional rankings. The data this week showed that the median enterprise B2B brand is cited in just 3% of relevant AI Overviews, and fewer than 26% of marketers track this at all. Tools for tracking AI citation rates across ChatGPT, Perplexity, and Google AI Mode are available now. Add this to your reporting this week, even as a manual audit, so you have a baseline to improve from.
- Audit your website's accessibility markup before your next development sprint. With 57.2% of web traffic from automated bots and 95.9% of homepages carrying detectable accessibility failures, this is no longer a compliance question. It is a visibility question. Book a developer accessibility audit and prioritise native HTML element use over ARIA patches. The dual benefit, better AI readability and better user experience, makes the investment straightforward to justify.
- Make every creative investment earn on two dimensions. Whether you are running Demand Gen, creator campaigns, paid social, or organic content, the evidence from Cannes, challenger brand research, and Google's own YouTube data all points the same way. Creative that builds the brand while driving immediate response outperforms creative optimised for only one of those outcomes. That principle applies at any budget level.
Want help putting any of this into action?
Whether it is AI search visibility, PPC performance, brand positioning for tougher trading conditions, or making sense of what these platform changes mean for your specific strategy, we can help.
Get in touch with The Digital Maze. 
