Most marketing predictions are designed to be forgotten. They are broad enough to be right and specific enough to feel credible without actually committing to anything. WPP’s latest AI advertising forecast is not one of those predictions.
It puts a number on the AI search advertising opportunity that is specific, uncomfortable, and backed by a company managing tens of billions in global media spend. By 2031, AI search advertising will account for 39.2% of the global search ad market. Today it sits at 1.9%.
That is not a long-range trend to monitor. That is a channel transformation happening inside your current planning horizon, while your 2027 budgets are still being written.
This week also brought news from Meta, Microsoft, Google, and Bing that collectively tell the same story from different angles: the infrastructure for AI-native marketing is being built right now. The decisions you make in the next six months about where to invest, what to measure, and how to maintain visibility will determine your position in that market.
Here is what mattered, why it mattered, and what to do with it on Monday.
Board Analysis is The Digital Maze’s weekly newsletter for UK marketing leaders. Every Wednesday, we cut through the noise and deliver the updates that actually affect your strategy. Subscribe here to get it straight to your inbox.
Search & Intelligence: Bing Has Given You the Metrics. Now You Need the Strategy.
Microsoft’s new Citation Share reporting in Bing Webmaster Tools is not a minor update. It is the first meaningful measurement tool for AI search visibility that does not require third-party interpretation or proprietary tooling. For the first time, you can see exactly what percentage of AI-generated citations your site receives for specific queries compared to all other cited sources, segmented by intent type, and tracked across time periods.
The timing matters because Ahrefs, in its analysis of 137,000 domains published this week, found that 97% of llms.txt files received zero requests from AI retrieval bots in May 2026. Of the files that did receive traffic, AI retrieval bots accounted for just 1.1% of total requests. The technical shortcut that many have adopted to signal AI accessibility is not being used. That is not an argument against good technical foundations. It is an argument for focusing on the fundamentals that already work: entity clarity, topical authority, and HTML structure that AI crawlers can actually parse without rendering JavaScript.
The strategic implication, as Search Engine Land explored this week in its analysis of how AI is merging paid and organic visibility, is that the signals from both channels now need to tell the same consistent story. Google’s Gemini infrastructure powers paid and organic results simultaneously. A brand’s organic authority directly affects paid Quality Scores. Inconsistency between what your organic content says and what your paid creative says creates what the article terms a “Doubt Tax” on overall visibility.
Why this matters:
- Citation Share in Bing Webmaster Tools shows your exact AI citation rate for specific queries, not just impressions. It is the measurement foundation that has been missing from AI search reporting, and it is freely available now.
- Ahrefs found that 97% of llms.txt files received zero AI requests in May 2026. The file format is not the solution. Content quality and accessible technical structure are what AI retrieval systems use.
- As AI merges paid and organic visibility under the same infrastructure, your brand’s organic authority directly affects your paid performance. The two channels are no longer independent, and strategy needs to reflect that.
Read more: “Bing Webmaster Tools updates AI reporting with Intents, Topics, Citation Share and Compare,” Search Engine Land
Also worth your time this week:
Paid Media & Performance: Three Google Ads Changes, One Announcement, a Hard August Deadline
Google has a well-established pattern of bundling significant bidding changes into announcements that feel operational until you read the detail. This week’s three-update release follows that pattern. Smart Bidding Exploration (SBE) is now globally available for Search and Performance Max campaigns across all languages. Promotion Mode enters beta for structured budget adjustments around promotional events. And Budget-Limited Campaign Optimisation rolls out on 17 August 2026, with account notifications arriving on 6 July 2026.
The August change is the one that requires action now. Google Ads Liaison Ginny Marvin confirmed the update will adjust backend bidding target optimisation for campaigns that are regularly constrained by budget. The system will change automatically. Campaigns with CPA and ROAS targets that have not been reviewed in light of current performance could see unexpected benchmark shifts when the change lands. Three weeks between notification and rollout is not a generous review window for accounts managing multiple budget-constrained campaigns.
The WPP AI advertising forecast adds strategic context here. If advertisers will be transferring traditional search and e-commerce budgets into AI ad options over the next five years, the accounts performing most efficiently today will be best placed to fund that transition. Every percentage point of ROAS improvement now is leverage for the strategic shift ahead.
Why this matters:
- Smart Bidding Exploration is now live for all languages on Search and Performance Max. Google expects “minor performance fluctuations” during calibration. On a high-spend account, minor fluctuations are not minor. Review active SBE settings today.
- The 17 August Budget-Limited Campaign Optimisation change will adjust bidding targets automatically for budget-constrained campaigns. Account notifications arrive 6 July. Review CPA and ROAS targets before those notifications land.
- Demand Gen billing is shifting from CPC to CPM for some Discover campaigns from 15 July 2026. Advertisers using view-through conversion optimisation should check whether this affects their spend pacing and can opt out by disabling VTC optimisation in campaign settings.
Read more: “Google Ads Bidding Changes: What PPC Managers Need To Know About The 3 Updates,” Search Engine Journal
Also worth your time this week:
Web Development & Design: Google Has Settled the HTML vs Markdown Debate
The trend of converting website HTML to Markdown as an AI optimisation shortcut has circulated in certain SEO circles this year. The argument runs that simpler content is easier for LLMs to parse. Google has pushed back on it directly this week, and the rebuttal is unambiguous. John Mueller’s position: “Converting HTML into text is trivial. There are lots of libraries out there that can do that for you.” What you strip away with Markdown is the architectural context that connects individual pages to the rest of the site.
Martin Splitt makes the technical case clearly. Navigation structures, internal links, and header hierarchies are the signals that tell search engines, and AI retrieval systems, how content relates across a site. That relational context cannot be recovered from plain text. Strip it out in the name of AI readability, and you are reducing the discoverability of every page you convert. The irony is that most major AI crawlers do not render JavaScript either, which means semantic HTML is often the only machine-readable version of a page available to them. Clean, structured HTML serves both Google ranking and AI retrieval simultaneously. Markdown serves neither.
The Schema.org usage dataset published this week provides the benchmarking tool that has been missing from structured data strategy. For the first time, you can see exactly how many domains across the public web are using each schema type, grouped into adoption range buckets and updated monthly. That is the context needed to evaluate whether your current schema implementation is competitive within your sector.
Why this matters:
- If your team has been converting pages to Markdown for AI optimisation, Mueller’s guidance is clear: HTML conversion is trivial for any crawler. What you lose is the site structure that search engines use to understand how your content connects. That is not recoverable.
- AI crawlers do not render JavaScript. Semantic HTML is often the only machine-readable version of your page available to AI retrieval systems right now. Your existing HTML structure is already serving that function, if you keep it intact.
- Schema.org’s new public usage dataset lets you benchmark your structured data implementation against real domain adoption data for the first time. Monthly updates make it a live competitive reference, not a static benchmark.
Read more: “Google Says Markdown For AI SEO Strips Away The Parts That Matter,” Search Engine Journal
Also worth your time this week:
Brand & Connection: The Recession Conversation Is Already Here. Are Your Campaigns Ready?
EY-Parthenon’s Consumer Sentiment Survey finds that nearly two-thirds of consumers believe a recession is likely. Financial confidence has declined 12% over six months. The Consumer Price Index rose 4.2% in the 12 months to May, with energy prices accounting for over 60% of the monthly CPI increase. Nearly one-third of consumers are reevaluating leisure and entertainment budgets, 29% are reconsidering food budgets, and one-quarter are changing travel plans. These are not background economic conditions. They are the context in which your next campaign will land.
The strategic implication is not to abandon aspirational positioning. It is to ensure that whatever your brand communicates connects clearly to value. EY-Parthenon’s Americas retail sector leader Will Auchincloss frames it plainly: “They want the brands to be on their side and authentic, they want them to be affordable, and they want them to be good value for what they’re purchasing.” Authenticity is not a brand direction to deploy selectively. During an economic downturn, it is a commercial requirement for brands that want to retain loyalty.
Arthur Sadoun’s pre-Cannes remarks add an adjacent point worth noting. His warning that the industry has been “overpromising on AI in the last three years” is not anti-AI scepticism. It is a reminder that the channel which earns trust during an economic downturn is not the most technologically sophisticated one. It is the most consistently human one. “The only answer is the proof,” Sadoun stated. That applies equally to AI claims and to brand value messaging right now.
Why this matters:
- Financial confidence declined 12% over six months. Campaigns built around pure aspiration without a credible value narrative will not convert in this environment. Value messaging needs to be built in, not bolted on.
- Consumers are rotating product categories and shopping channels to save money. Understanding which channel they are moving to, not just that they are moving, is the planning question that should be live in your strategy now.
- Sadoun’s Cannes framing that “the only answer is the proof” applies across channels. Brands that can demonstrate specific, tangible value to consumers right now will outperform those relying on positioning statements and reach metrics.
Read more: “Consumers are worried about a recession. How should brands respond?”, Marketing Dive
Also worth your time this week:
AI & Automation: Meta Makes Community Conversations Into Search Results. Taboola Builds the Ad Layer.
Meta has introduced AI Mode to Facebook Search. Unlike Google’s AI Overviews or Bing’s AI answers, which draw from the open web, Meta’s AI Mode generates responses from public content within its own ecosystem: Facebook Groups, Reels, and other Meta properties. It uses Meta AI and Muse Spark to surface what Meta describes as “real answers from real people.” The commercial and brand implication is immediate.
If your brand, product, or service is discussed in public Facebook Groups, those conversations are now direct inputs to how you appear in Facebook Search AI results. You do not own that content. You cannot edit it. You cannot request its removal from the ranking system. And Meta has not disclosed how content is selected or weighted within AI responses, meaning there is no optimisation path available. What this makes clear is that community management, authentic brand presence in public Group spaces, and the quality of the organic conversations happening around your brand are no longer soft metrics. They are a direct input to AI search visibility on the world’s largest social platform.
Taboola’s expansion of DeeperDive into a full ad network for LLMs, chatbots, and virtual assistants, announced this week, is the commercial infrastructure layer for exactly this kind of AI-native discovery. Third-party AI platforms integrate Taboola’s native ads into conversational interfaces, advertisers pay cost-per-click, and publishers take 65% of revenue. Publishers using DeeperDive are seeing a 10% monetisation lift on existing revenues. The economic model for advertising inside AI conversations is being established now, at a stage when first-mover advantage is real and entry costs are lower than they will be at scale.
Why this matters:
- Facebook Search now surfaces AI responses powered by public Group conversations and Reels. Community management and authentic brand presence in public Groups are direct inputs to Meta AI search visibility. This is not a future consideration.
- Meta has not disclosed source selection criteria for AI Mode responses. There is no optimisation path. The only lever available to brands is the quality and authenticity of the public conversations already happening around them on the platform.
- Taboola’s DeeperDive expansion shows the ad network infrastructure for AI conversational interfaces is being built now. Publishers are seeing 10% monetisation lift. Early test investment in AI ad formats will be cheaper than buying market position at scale.
Read more: “Meta launches AI Mode in Facebook search to answer questions,” Search Engine Land
Also worth your time this week:
Case Study of the Week: 694% Revenue Growth on a Limited Budget
Elsons Tools is a family-owned UK supplier of tools, building materials, and workwear, established in 1968. When they came to The Digital Maze, their Google Ads account was underperforming. Tight profit margins, a large product range, and a limited budget meant they needed real results fast, against a 3-5% ROAS target while competing against national retailers.
We restructured the account with inventory filters to prevent products competing across multiple campaigns, worked with the client to exclude low-margin and non-priority product categories, and transitioned to smart bidding with a target ROAS strategy focused exclusively on competitively priced, well-stocked items.
The results in Q1, year-on-year: +694% increase in revenue, +439% increase in ROAS, consistent performance above the 3-5% ROAS target for over eleven consecutive months, and Silver at the UK Search Awards for Best Budget PPC Campaign.
“Web sales have grown significantly since they came on board. Experienced team and fairly priced.” Jon Valentine, Elsons Tools.
Read the full Elsons Tools case study.
Bringing It Together
This week’s news is structurally consistent. AI is expanding into advertising, search, and social discovery simultaneously. The measurement tools are arriving. The forecasts are hardening into numbers. The ad network infrastructure is being built.
The marketing leaders who will be best positioned in 2031 are not those who wait until AI advertising reaches 10% or 20% of the market before committing to a view. They are the ones who have already tested, measured, and understood what works for their brand, their audience, and their category before budgets are at scale.
At the same time, the fundamentals have not changed. Clean technical foundations. Specific, credible messaging. Value propositions that hold up under economic pressure. Audiences who trust you because you have been on their side consistently. Those are the assets that compound in any channel, including ones that do not exist yet.
Three Takeaways for the Week Ahead
- Get into Bing Webmaster Tools and review your Citation Share data this week. If you have never looked at your AI citation performance in Bing’s new reporting, this is the moment. Citation Share is the most actionable AI visibility metric available right now, it is free, and it gives you a baseline before the channel grows further.
- Review budget-limited Google Ads campaigns before 6 July. Account notifications about the August bidding changes arrive on 6 July. Identify which campaigns are regularly constrained by budget, review current CPA and ROAS targets, and decide whether any need adjusting before the automated changes go live on 17 August. Do not wait for the notification to act.
- Build your recession-resilience messaging now, not when conditions worsen. Nearly two-thirds of consumers already believe a recession is likely. Brands that have a clear, credible value narrative ready to activate will outperform those who scramble to adapt when consumer confidence falls further.
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.
