July showed search becoming more visual, more social and more agentic. Google brought AI image generation into AI Overviews, added social post reporting to Search Console and expanded AI Mode with connected apps that can help users complete tasks without leaving the search experience. At the same time, the rules governing AI access to the web are becoming more defined. Cloudflare introduced granular controls that let site owners manage AI crawlers by purpose – Search, Agent and Training – while Google and Anthropic continued pushing agents deeper into everyday discovery and decision-making. The direction is clear: visibility now depends on being accessible, trustworthy and useful across a much wider ecosystem than the traditional search results page.
Here’s what changed across Google, AI, Agentic, Social, and Paid and what it all means for brands navigating the next phase of search.
Google Search
Google Adds AI Image Generation to AI Overviews and Redesigns Google Images Homepage
Marking 25 years of Google Images, Google is bringing image generation directly into AI Overviews using its Nano Banana model, turning a text prompt into a custom visual when no suitable image exists on the web. It is also redesigning the Google Images homepage into a personalised, real-time gallery with saveable collections, a clear nod to Pinterest-style discovery. Image generation rolls out over the coming weeks in English for regions already supporting image creation in AI Mode, with the Images redesign starting on US desktop.
Why It Matters? This extends the answer-layer dynamic from text to visuals: if users can generate the image they need inside the results page, that is one more reason not to click through to a publisher or stock site. Both moves point the same way, with Google rebuilding its core products around retention rather than referral. For clients, it reinforces that distinctive visual assets and genuinely useful content matter more than ever, since generic imagery is now the easiest thing for Google to replace in-house.
Google Search Console Adds Reports For Social Posts
Google is introducing a new property type in Search Console called platform properties, which lets you monitor how social media and video posts perform in Google Search and Discover. This feature supports platforms such as Instagram, TikTok, X, and YouTube and is available even to creators who don’t have their own websites.
Why It Matters? This indicates how omni-channel search is becoming – Google Search Console is usually reserved for websites. This function will allow account holders to see data across platforms – not just Google.
Google Says Search Hit an All-Time Usage High During the World Cup
Google’s Nick Fox and Robby Stein both confirmed on X that Google Search usage hit an all-time record right after Argentina’s stoppage-time comeback win over Egypt on 7 July. No figures were shared, and Google hasn’t published a blog post or methodology behind the claim. Top searches included “argentina vs egypt” and “how many world cup goals does messi have.
Why It Matters? Google has made near-identical claims before, including after the 2022 World Cup final, and it’s notable that this comes as the company faces persistent narrative pressure about AI chatbots eroding traditional search. Worth treating with some caution for reporting purposes: record usage is an internal query-volume metric, not a proxy for clicks to publisher sites, and Google itself hasn’t disclosed how much of that volume converts to outbound traffic.
Google Brings Connected Apps to AI Mode in Search
Google announced on 16 July that users can connect selected third-party apps directly to AI Mode in Search. The initial rollout includes services such as Instacart, Canva and YouTube Music. This allows users to interact with connected services within a Search conversation rather than moving between individual apps. The feature began rolling out in the US during the week of the announcement, with more partners expected over time.
Why It Matters? AI search is moving beyond answering questions toward helping users complete tasks across connected services. This can reshape product discovery, planning and conversion journeys, with fewer hand-offs between Search and external platforms. For brands, it raises the strategic value of being present in the data sources, apps and platforms AI assistants can access.
It also reinforces that search visibility is increasingly connected to ecosystem partnerships, structured data and useful digital experiences, not only website rankings
AI Search
OpenAI Launches GPT-Live, Bringing Web Search Into ChatGPT Voice
On 8 July, OpenAI began rolling out GPT-Live, a new full-duplex voice model that can listen and speak at once and now powers ChatGPT Voice for over 150 million weekly users. When a spoken question needs current information, GPT-Live hands it to a frontier model (GPT-5.5 at launch) to run a web search, then weaves the answer back into the conversation. It also surfaces visual cards for weather, stocks, sports and maps mid-conversation. OpenAI has not detailed how it handles source citations for those spoken answers.
Why It Matters? Voice is becoming a genuine discovery surface, and one where a searcher may get a complete answer without ever seeing a link. For clients, that raises the stakes on structured, accurate, well-cited content and clean product and location data, since GPT-Live pulls from what the model already trusts. Thin or unsourced pages risk being skipped entirely in a spoken comparison. The missing citation trail is the open question worth watching for measurement and attribution.
New AI traffic Options Released for Customers on CloudFlare
Cloudflare has introduced new AI bot traffic controls that allow all customers, including those on its free tier, to manage AI bot traffic by three distinct use cases instead of a blanket block: Search (indexing content to answer queries), Agent (bots acting on a user’s behalf in real time), and Training (crawling to build/fine-tune models). Starting September 15, 2026, new domains onboarding to Cloudflare will have Training and Agent bots blocked by default on ad-monetized pages, while Search remains allowed by default, since ads signal that human attention is the monetizable goal the site wants to protect.
Why It Matters? Cloudflare’s taxonomy forces bot operators to separate these functions, meaning AI search products may need distinct, clearly declared crawlers if they want reliable content access. Default blocking on ad-monetized pages could reduce the pool of crawlable content for AI training and agentic tools, while explicitly preserving access for Search-classified crawlers. Since Cloudflare sits in front of a large share of the web, these defaults will materially affect what content AI crawlers – including major search/agent bots – can actually retrieve, making this a structural shift in the data supply chain that AI search products rely on, not just a policy statement.
Google AI Mode’s shift to citing itself
Between April 15 and June 30, 2026, google.com’s citation share in Google’s AI Mode surged 8.4x, making it the second most-cited domain, driven almost entirely by Google’s own Business Profile and Product Knowledge Panel subpaths. Google is now presenting these as inline, interactive “popup panels” that function as a brand’s primary landing page, bypassing the traditional click-through to a company’s own website. This effect is concentrated in local-intent industries – hospitality/travel, home services, restaurants, real estate, and healthcare – and in product comparison queries, where Product Knowledge Panels are shown instead of a direct link to a retailer’s site.
Why It Matters? This is a concrete signal that Google is trying to keep users inside its own ecosystem rather than routing them to brand websites – turning AI Mode into a “zero-click” layer where Google-hosted cards, not your homepage, become the actual point of conversion. Google Business Profile hygiene, structured product data, and review sentiment become the new ranking surface. It’s also an early glimpse of Google’s push toward agentic, zero-click commerce, suggesting the gap between “being cited” and “being clicked” will keep widening in AI-driven search.
Google Merchant Center AI Performance Insights For AI Mode & AI Overviews
Google’s Merchant Center AI Performance Insights report, first announced at Google Marketing Live in May, is now live in a pilot with a limited number of US Merchant Center accounts as of 14 July. It sits under Analytics → Products → AI performance and reports Share of Voice, competitors’ average share, and query type/frequency, benchmarking a brand’s visibility in AI Mode, AI Overviews and the Gemini app against similar brands. Google frames it as helping merchants understand the shift “from keyword-based searches to complex, conversational shopping paths.” It’s organic-only, impression-based, and has no clicks metric.
Why It Matters? This is the first native report Google has given merchants specifically for AI shopping surface visibility, but access is still pilot-only and US-only for now (Australia, Canada, India and New Zealand are promised “in the coming months,” no date given). It’s also drawing attention from leading voices in search, who flagged it as one of the first real windows into how conversational shopping queries group and behave, rather than just whether a brand gets cited.
Agentic Search
Claude Cowork Expands to Web and Mobile
On 7 July, Anthropic expanded Claude Cowork, its agent for general knowledge work, from desktop to web and mobile, starting with Max subscribers. Sessions now run in the cloud, so a task started at your desk continues in the background and scheduled tasks run with no device online. Anthropic’s own data from 1.2 million sessions showed over 90% of Cowork use is not coding, led by business operations (33.4%) and content creation (16.4%). Chat and Cowork were also merged into a single interface.
Why It Matters? Agents are moving from developer tools into the everyday marketing and ops surfaces where our work actually happens, from drafting decks to reconciling data to content production. The scheduled, background-running model is the notable shift: an agent can prepare a client briefing overnight and surface only the decisions that need a human. For agencies, this reframes how routine deliverables get produced and reviewed, and it is worth trialling on repetitive reporting tasks first.
Anthropic: Redeploying Claude Fable 5
Anthropic announced it is redeploying Claude Fable 5 globally starting July 1, after the US Department of Commerce lifted the export controls that had prompted Anthropic to suspend access to Fable 5 and Mythos 5 on June 12. The redeployed model ships with a new set of classifiers built with US government input to specifically target and block cybersecurity misuse, following nearly ten weeks of engagement with the Office of the National Cyber Director, the Office of Science and Technology Policy, Treasury, Commerce (including CAISI), and national security agencies. Anthropic also announced a new HackerOne program for researchers to report cyber jailbreaks, and said it’s drafting a consensus framework with Amazon, Microsoft, Google and other Glasswing partners for assessing jailbreak severity industry-wide.
Why It Matters? Teams already relying on Claude for AI-assisted content, research or dev support on entry-level plans should budget for this now rather than getting caught out mid-project: a $100 credit doesn’t stretch far at these token rates, and the gap between what a Pro seat and a Max seat can now do has widened noticeably. It’s also a useful data point for any client conversations about AI tooling costs, since it shows premium model access increasingly tracking premium subscription tiers rather than being bundled in as standard, a pattern likely to repeat across other providers as frontier models get more expensive to run.
Google Brings Agentic Experiences to Search
Google has announced information agents for Search that can monitor the web and real-time information for a user’s specified needs. Instead of requiring repeated searches, these agents can send updates when relevant changes occur. Google is also expanding agentic experiences that help users compare availability and pricing for local services. Selected Search experiences will be able to contact businesses on a user’s behalf, while users complete bookings with the provider.
Why it matters? This marks a shift from search as a one-off retrieval activity to search as a delegated, ongoing task. Brands must make service information, availability, pricing and transactional paths accurate and easily accessible to AI systems. For SEO, being discoverable will increasingly mean supplying trustworthy, structured content that supports an agent’s research and recommendations. The opportunity is not just to rank for a query, but to become the reliable source an agent uses to complete a user’s task.
Social Search
TikTok Expands AI Content Labelling and Launches In-App AI Literacy Hub
TikTok’s official newsroom announced three connected moves: expanding AI literacy resources with media literacy group NAMLE and researcher Henry Ajder, testing improved detection for accounts dedicated to posting AI-generated spam, and building an in-app hub where users can learn to identify AI-generated content when they search AI-related terms. TikTok said it removed more than 86 million fake accounts in Q1 2026 alone.
Why It Matters? This ties directly into how people search and evaluate content on TikTok, an in-app hub surfaced via search terms is a new discovery surface in its own right, and stricter detection of AI spam accounts could reshape which content survives in search results. Worth flagging to any client whose organic TikTok strategy leans on AI-assisted content production, since enforcement details (account-level thresholds, what counts as “spam”) haven’t been published yet.
YouTube Clarifies Policies Around AI Slop and Upsetting Videos
YouTube rolled out clarified guidelines on 16 July defining three types of “inauthentic content” under its existing YouTube Partner Program policy that cannot be monetised: generic or repetitive template-style videos, content that’s “unsatisfying or off-putting” (such as distressing footage where that distress is the point of the video), and videos featuring fake AI-generated “experts.” YouTube’s Conor Kavanagh said any channel with too much of any of these three categories will lose monetization eligibility. The policy itself isn’t new, this is a clarity update giving creators and reviewers clearer, more actionable language for why a channel gets flagged.
Why It Matters? This sharpens the line between AI-assisted content that still adds editorial value and mass-produced “slop,” right as more brands lean on AI tools for video production at scale. Any client running or working with YouTube channels for organic or influencer content should sense-check current output against these three categories now, since demonetisation risk sits with the channel, not just the individual video, and the guidance leaves real ambiguity around what counts as “off-putting” for news, true crime or reaction-style content.
Google Images gets a Pinterest-like redesign focused on discovery
Google is redesigning Images into a dynamic, personalised “For You” gallery rather than a purely query-led results page.The feed is tailored to users’ interests and browsing history, refreshes with new web imagery, and lets people save ideas into collections. Google is also adding AI image generation in Search and AI Overviews for visual concepts that do not already exist online. The initial rollout is scheduled for English-language desktop users in the US, with further expansion expected.
Why it Matters? Visual search is moving closer to social-feed behaviour: inspiration can now be surfaced before a user enters a keyword.For fashion, fitness and e-commerce brands, image quality, descriptive alt text, product structured data and consistently distinctive visual assets become even more important. The change increases the value of publishing “save-worthy” content such as styling ideas, product comparisons, workout inspiration and creator-led imagery. It also makes Google Images a more direct competitor to Pinterest for high-intent discovery journeys.
EU challenges Meta’s engagement design
The European Commission preliminarily finds Meta’s “addictive” Instagram and Facebook design may breach the DSA. The European Commission has preliminarily found that Meta may be breaching the Digital Services Act through the addictive design of Facebook and Instagram.Its investigation focuses on infinite scroll, autoplay, push notifications and highly personalised recommendation systems. The Commission says Meta did not adequately assess the wellbeing risks to users, including minors and vulnerable adults.
It also says Meta’s current mitigations did not effectively address those risks; these are preliminary findings, not a final decision
Why it Matters? If upheld, the case could force changes to core feed and recommendation mechanics that currently drive discovery and engagement on Meta platforms. That could affect content distribution, watch-time expectations, notification strategies and the metrics marketers prioritise.Brands should reduce reliance on passive-scroll reach by building stronger direct community signals, searchable evergreen content and owned audiences. The story also raises the strategic importance of creating content that earns intentional search and saves, rather than only algorithmic interruption.
Paid Search
Meta Expands AI Personalisation & Video Ad Capabilities
Meta has expanded AI-powered personalisation across its advertising ecosystem while rolling out Live Video Ads to Instagram globally. The updates are designed to improve campaign relevance, increase engagement, and give advertisers more opportunities to reach audiences with immersive video formats across Facebook and Instagram.
Why It Matters? Video continues to dominate digital advertising, and Meta’s investment in AI-driven personalisation means advertisers can deliver more relevant creative at scale. Brands that embrace these capabilities may see improved campaign performance while reducing manual optimisation efforts.
Retail Media & Connected TV Continue Rapid Growth
Walmart’s acquisition of CTV advertising platform Vibe.co highlights the growing importance of retail media and connected TV advertising. The deal strengthens Walmart’s ambitions to compete more directly with Google, Meta, and Amazon by making self-service CTV advertising more accessible for brands of all sizes.
Why It Matters? Retail media and CTV remain two of the fastest growing digital advertising channels. As more commerce and audience data become integrated into advertising platforms, marketers will gain better targeting capabilities and clearer measurement across the customer journey. Brands that diversify beyond traditional paid search and social will be well positioned for future growth.
TikTok Expands AI Advertising Tools with Symphony
TikTok has expanded its AI-powered creative suite, Symphony, introducing new tools that help advertisers generate, edit and optimise video creatives at scale. The update also includes enhancements to Smart+, TikTok’s automated campaign solution, with AI selecting the best performing creative assets and providing campaign optimisation recommendations. These features are designed to reduce production time while improving campaign performance.
Why It Matters? Creative remains the biggest driver of performance on TikTok, and Symphony makes it easier for brands to produce high quality, platform native content without significantly increasing production costs. For paid media teams, this means faster creative testing, improved campaign efficiency, and greater reliance on AI to optimise performance allowing marketers to focus more on strategy and less on manual execution.
What Happened Last Month
In June, AI search moved beyond retrieval into action. Google doubled down on zero-click search, AI assistants became capable of shopping, booking and researching on users’ behalf, and platforms introduced the first real tools for measuring AI visibility. At the same time, social platforms strengthened their role as AI search sources, while paid media expanded further into conversational experiences with ChatGPT ads rolling out internationally and Google continuing to reshape advertising for AI-powered search. Across every channel, the focus is shifting from generating traffic to understanding influence, as AI continues to reshape how visibility is measured and valued.
Read the full June Search Update here.
