Every Friday, we publish the week’s most important developments across AI, search, and digital marketing. It’s the public version of the intelligence brief our team reads to run client work: the sources we actually watch, plus our take on what each move means for a business owner. Curated by Spencer K. Wade and assembled in the Lift AI Lab.
Prefer to listen? An audio version of this week’s brief is available: listen on the go.
The 30-second version
- The EU ordered Google to share its search data with rivals starting in 2027, and AI chatbots are explicitly included. The data advantage that made “SEO” and “Google” synonyms for twenty years is being pried open on a dated timeline. 2027 is one budget cycle away.
- AI Mode ads now appear on roughly 30 percent of queries, and paid placement rarely buys you a citation. You cannot purchase your way into the answer. Being cited and being advertised are now two separate jobs.
- Google is the second most-cited domain inside its own AI Mode. Business profile and product feed data are becoming citation assets, not just listing assets, especially for local and retail businesses.
What deserves your attention this week
1. The EU is forcing Google to share search data, and AI chatbots get access too. Under the ruling, Google must share anonymized search data with traditional search engines and with AI chatbots that have search functionality, subject to privacy safeguards. Qualifying chatbots can access the same data Google uses to improve its own services. A separate EU decision is expected on whether Google favored its own services in search, with required changes that could reshape how comparison, travel, and shopping platforms surface on high-value queries. Here is why it matters to you: Google’s search quality has been underwritten by a data asset nobody else could match. If Perplexity, ChatGPT, and others get comparable query data in 2027, the quality gap between AI answer engines and traditional search narrows structurally. Visibility inside AI answers stops being a speculative bet and starts being infrastructure with a due date on it. (Source: Search Engine Land)
2. AI Mode ads hit about 30 percent of queries, and paid does not buy citations. A new study puts AI Mode ad coverage at nearly 30 percent of queries, with high-CPC commercial searches far more likely to show AI text ads. The finding that matters most is what did not happen: paid placements rarely boosted citations or organic rankings. That is an empirical answer to the most common question we get, which is some version of “can I just buy my way into the AI answer?” No. The answer layer decides who it trusts on different evidence than the ad auction uses, and that means visibility work and paid work are complementary budgets, not substitutes. (Source: Search Engine Land)
AI and models this week
Google is citing itself inside AI Mode. Google-owned business and product cards are taking a growing share of AI Mode citations, particularly on local and shopping queries. In parallel, AI Mode added Instacart, Canva, and YouTube Music integrations for US users, letting people move from planning straight to checkout, design, or listening without leaving Search. Two things follow. If you are a local or retail business, your Google Business Profile and Merchant Center feed hygiene are now citation inputs, not just directory chores. And the direction of travel is clear: the answer surface wants to complete the transaction, not hand it off. (Source: Search Engine Land)
A second “DeepSeek moment,” and the panic is probably overcalibrated. Moonshot released Kimi K3 with plans to publish the weights, timed to the World AI Conference in Shanghai, and one outlet declared America’s AI lead erased. A more careful read cuts against the hype: by Moonshot’s own admission, K3 is not at the frontier. Open-weighting a model that is not at the frontier is a low-risk, high-leverage move for a country that is behind, and the expectation is that once genuinely frontier-level capability arrives, the open releases stop. For a business owner, the practical read is narrower than the headlines: capable, inexpensive models remain widely available for now, but that is a window, not a permanent condition. Build so you can swap what is underneath. (Source: Transformer, The Neuron)
Compute and power constraints hardened on three fronts. New York imposed a one-year moratorium on new data center construction over 50 megawatts. Chip export control bills are reportedly headed into the Senate defense bill. A major memory supplier signalled the shortage runs to 2030. Everyone has been assuming AI costs only fall. If you are budgeting an AI-powered product or workflow with per-query economics, price in the possibility that inference costs go flat or up through 2027 rather than assuming the usual deflation curve. (Source: Transformer Weekly, The Neuron)
Search, SEO, and the AI-search layer
AI visibility finally has a defensible ROI framework. The argument, from Dale Bertrand: citation share, rankings, and impressions are operational metrics. They become business metrics only when tied to revenue opportunity, revenue at risk, pipeline influence, or customer acquisition cost. He pairs it with a method for estimating impact when attribution is incomplete. In one example, a healthcare company used its own sales-call data to estimate that a competitor’s messaging was influencing roughly $12 million in annualized pipeline. Alongside it, a practical audit model emerged: test the discovery, comparison, trust, and logistics questions your customers actually ask across ChatGPT, Perplexity, Gemini, and AI Overviews, then sort what you find into three buckets. Invisible, inaccurate, or misframed. Each bucket has a different fix, and the order matters. Fix crawl access and business data before you spend a dollar on new content. (Source: Search Engine Land)
The real commerce shift runs on data, not creative. The industry conversation has fixated on ChatGPT ads, which misses the larger move: AI commerce is heading toward agents that recommend and complete purchases, and in that world your product data outweighs your ad creative. The B2B version of this finding is sharper. AI agents get stuck on opaque pricing and poor machine readability, and pricing pages generated most third-party citations because the agents could not retrieve or verify the information from the source, even when the pricing was public. The fixes are unglamorous and cheap: crawlable HTML pricing instead of a JavaScript calculator, Product and Offer schema, AI crawler access, one canonical URL. If your pricing lives behind a form, a directory is very likely getting cited in your place. (Source: Search Engine Land)
Google Ads had a messy week: reporting inflation, new controls, and a bidding backlash. Three things landed together. Performance Max product reporting expanded across all eligible networks, which produces a one-time jump in reported metrics that is not a real gain. Local Inventory Ads are flipping to default-on for eligible Shopping campaigns. And a Smart Bidding update drew sharp criticism across the PPC community. The first one is the trap. If your July report shows a sudden improvement in Performance Max and nobody annotated the change, you are being handed a phantom win that will read as a decline next month. Ask whoever runs your accounts whether the break is annotated. (Source: Search Engine Land, SE Roundtable)
Attribution orthodoxy took three hits in one week. Marcus Miller argued the SEO versus PPC debate is over, because organic visibility strengthens paid, AI citations shape click behavior, and optimized content feeds AI-driven ads. Benjamin Wenner showed paid search performs better when paid social is running, so evaluating them in isolation leads you to overfund search and underinvest in demand generation. Ashish Jacob identified a fourth kind of content decay: zero-click capture, where your rankings hold steady but clicks fall, which means the answer layer is absorbing the traffic rather than your content going stale. Refreshing the page will not fix it, because the page is not the problem. All three say the same thing from different angles. Last-click reporting is now actively misleading, not merely incomplete. (Source: Search Engine Land)
Quick hits
- Apple Maps banned home services listings. If you are a contractor, plumber, or in any home services vertical, check your presence there. (SE Land)
- ChatGPT Ads audience lists are rolling out. The channel keeps maturing on a fast clock. (SE Land)
- Search Console added reporting on social and video platforms. (SE Land)
- AI labels on Google Ads are rolling out, with AI ad disclosures expanding across Search, YouTube, and Discover. (SE Land)
- Canonicalization fixes can take up to two weeks to resolve, per Google. Useful for setting expectations after a site change. (SE Roundtable)
- Merchant Center added an AI performance report. (SE Land)
- The Job Indexing API is weaker than assumed. An HTTP 200 response does not mean indexed, and quota requests can go unanswered for months. (SE Land)
- Bing confirmed it does not issue one-off search penalties. (SE Roundtable)
What we’re advising clients this week
Annotate the Performance Max reporting break now. The metric jump is a reporting change, not performance. Flag it in this month’s report so next month’s correction does not read as a loss.
Audit your Local Inventory Ads and bid strategy settings this week. Local Inventory Ads are going default-on for eligible Shopping campaigns and the Smart Bidding update is live and contested. Both can shift spend somewhere nobody chose.
Stop measuring AI visibility as one number. Test the questions your customers actually ask, across the assistants they actually use, then sort what you find into invisible, inaccurate, or misframed. Those three problems have three different fixes and doing them out of order wastes money.
Tie AI visibility to dollars or stop reporting on it. Citation share is an operational metric. Revenue opportunity, revenue at risk, and pipeline influence are business metrics. If your agency cannot make that translation, you are buying activity.
Make your pricing and product data machine readable. Crawlable HTML pricing, Product and Offer schema, AI crawler access, one canonical URL. This is the cheapest visibility work available right now and it is where agents are currently failing.
Form your point of view on the 2027 data-sharing change. It moves AI-surface visibility from speculative to scheduled. Budgeting for it a year late is the expensive version.
Sources
Search Engine Land, Search Engine Roundtable, Transformer Weekly, and The Neuron, with commentary from Dale Bertrand, Adam Heitzman, David Carrasco Pamies, Rémi Kerhoas, Kevin Indig, Marcus Miller, Benjamin Wenner, and Ashish Jacob. Synthesized from 21 newsletters for the week of July 10 to July 17, 2026. Figures such as AI Mode ad coverage and pipeline influence estimates are as reported by the cited studies and are worth validating against your own analytics before acting.
This is a weekly series from Lift Conversions. We read the industry so you don’t have to, and translate it into what actually matters for your business.
Want to talk through what any of this means for you? Book a 30-minute call.
Related Posts
