AI Marketing News, 24 Aug 2026: Reddit Citations Down 86%, ChatGPT Ads in Europe, Microsoft AI Max Global
Reddit's ChatGPT citation share fell 86% in four days, ChatGPT Ads reach 31 European markets, Microsoft AI Max rolls out globally for Search.
Reddit’s share of ChatGPT Search citations fell 86.4% in four days. Promptwatch, reported by Search Engine Land on August 19, tracked Reddit at an average 3.83% of ChatGPT Search citations between July 18 and August 7. The share slipped from the high 3% range into the mid-2% range on August 8, the same day Promptwatch says ChatGPT Search changed its query fan-out behaviour, then fell below 1% on August 14 and averaged 0.52% through August 17. Google’s products showed nothing comparable: Reddit’s share in AI Overviews drifted from roughly 2.5% in early July to about 2.1% in August, and AI Mode declined from late July. Promptwatch says the data shows when the shift happened, not why, and cannot rule out a collection issue.
Importance for marketers: A retrieval change nobody announced cut one source’s visibility by 86% in four days. Your Reddit-thread strategy is a position in someone else’s index. Spread the sources a model can pull from, and treat any single-platform citation share as a number that can go to zero without warning. The durable version is the owned-surface layer described in citation infrastructure.
iPullRank panel data shows ChatGPT taking more paid clicks out of Google than any other top destination. The zero-click study, covered on August 17, analysed 13.1 billion search events from 9.1 million opted-in users between October 2024 and December 2025. ChatGPT ranked sixth among the most-clicked destinations from Google Search, behind YouTube, Google’s own properties, Reddit, Facebook and Wikipedia, but drew a higher proportion of paid clicks than any other leading destination in the set. Only 11.1% of searches heading toward OpenAI were intercepted by a zero-click Google experience, among the lowest rates measured. Across the period Google’s zero-click rate rose about 2.6 points and organic-click share fell about 2.8 points, while paid clicks stayed flat.
Importance for marketers: Zero-click growth in this dataset came out of organic, not paid, which inverts how most 2026 budget decks read. The paid line held because advertisers, OpenAI among them, kept buying their way through the answer layer. Build your AI-search plan as a content plan only and you fund the channel that took the loss.
ChatGPT Ads expands to 31 European countries. OpenAI announced on August 19 that ads reach 31 European markets the following week, its largest geographic expansion since the US test began in February 2026 and eight further markets after it. Germany, France, Spain, Italy, Sweden, Norway, Denmark, the Netherlands and Austria are named. Buying runs through OpenAI’s Ads Solutions team, agency partners and technology partners first, with self-service Ads Manager access expected later in the summer. Ads serve only to Free and Go users; Plus, Pro and Enterprise stay ad-free. OpenAI says tens of thousands of marketers have now advertised on ChatGPT, and lists conversion optimisation beyond CPM and CPC bidding, geo-targeting, custom audiences, the OpenAI Pixel, a Conversions API and third-party measurement integrations as platform additions.
Importance for marketers: The Conversions API and Pixel change your build; the geography does not. Server-side conversion posting means ChatGPT Ads sits in the same n8n or warehouse job that already posts offline conversions to Google and LinkedIn, so a European pilot runs on measurement plumbing you have. Wire the pixel and CAPI before self-service opens. Teams who start when the button appears spend three months rebuilding attribution they could have inherited.
Microsoft Advertising begins the global rollout of AI Max for Search campaigns. Product liaison Navah Hopkins confirmed on August 19 that AI Max is reaching all Microsoft Advertising accounts worldwide, opt-in for new and existing campaigns. The suite has three parts: search term matching, which goes beyond the keyword list using ads, landing pages and contextual intent signals including conversational queries in Bing and Copilot; text customisation, which generates messaging variations from existing assets and site content and picks combinations at auction; and final URL expansion, which selects a landing page matched to intent. Advertisers already running autogenerated text assets or predictive matching will find those settings migrated under AI Max, with the remaining features left off. AI Max settings carry over on Google Ads import, except campaigns that came from upgraded Dynamic Search Ads, which revert to DSA on Microsoft.
Importance for marketers: Gate final URL expansion. You are handing landing page selection to a model that reads your whole site, including the pages sales asked you to keep out of paid traffic and the ones legal has not cleared for claims. Write your URL rules before you switch it on, then check what your last Google Ads import turned on without asking you.
Meta AI can now connect to Meta Ads and Google Workspace to analyse and optimise campaigns. Meta announced the feature on August 20 for small businesses across web, mobile and desktop. Advertisers opt into connecting their Meta ad campaigns plus Gmail, Docs, Sheets and Slides to the assistant. Meta says it will identify which audiences are delivering, surface patterns across top-performing creative, flag ads that have stopped resonating and point out where budgets could work harder, with output as decks, documents and spreadsheets. Recurring tasks and reminders turn one-off analysis into scheduled reporting. Meta published no usage figures, pricing, rollout dates or performance benchmarks.
Importance for marketers: Meta is advising you how to spend on Meta, and it wants to read your Gmail and Sheets to do it. Your security review will stop at the second half. Run Meta’s recommendations against your own incrementality and profitability numbers before you act on them, which is the reconciliation layer any paid media agent needs and this product does not include.
Google’s A2A agent protocol moves to the Agentic AI Foundation, joining MCP. Axios reported on August 17 that Agent2Agent, the standard Google built for independent agents to talk to each other, becomes a hosted project of the Agentic AI Foundation, putting it under the same neutral body as Anthropic’s Model Context Protocol. A2A covers agent-to-agent communication; MCP covers connecting AI applications to tools and data. The foundation launched in December 2025 with fewer than 40 members and says it now has more than 250, including Google, Microsoft, Amazon, Anthropic, OpenAI, Bloomberg, Shopify and Block. Google Cloud VP Rao Surapaneni framed the original hypothesis as customers deploying agents from multiple providers that all need to work together. Axios gave no transfer date, and the foundation published no adoption numbers.
Importance for marketers: Put both protocols under one body and you get a plausible path to an enrichment agent from Clay, a CRM agent from HubSpot and an outbound agent from somewhere else coordinating without custom glue between each pair. It also weakens the argument for buying one vendor’s entire agent suite to get interoperability. Foundations move slower than product teams, so price this into next year’s renewals rather than this quarter’s build.
OpenAI previews zero-retention safety processing as Anthropic requires 30-day logs on top models. Axios reported on August 19 that OpenAI is testing Private Safety Processing, a system it says detects misuse spanning multiple interactions without requiring enterprise customers to give up zero-data-retention terms, with data either staying on customer-controlled infrastructure or held by OpenAI under customer-controlled encryption keys. Anthropic has gone the other way, requiring 30-day retention for customers using its most capable models on the argument that sophisticated multi-request attacks cannot be caught any other way. Both positions answer the same problem, and the split is now a visible disagreement between the two labs.
Importance for marketers: Retention terms decide which model can touch which data. If your enrichment or research agents process account lists, call transcripts or CRM records, a 30-day requirement puts customer data in a vendor’s logs for a month, which means a DPA amendment and a security review rather than a settings change. Check which of your agent workflows already route through a model that logs.
Pew: 52% of US adults are now more concerned than excited about AI. The Register reported on August 19 that Pew Research Center data shows public sentiment continuing to deteriorate, with the share primarily excited about AI falling from 18% in 2021 to 9%. Employment drives most of it: 71% expect AI to reduce US jobs over the next two decades, up from 64% two years earlier. The shift is sharpest among adults aged 18 to 29, 55% of whom are now predominantly worried, with many also expecting negative effects on society and creativity.
Importance for marketers: With a majority-negative audience, “powered by AI” costs you more than it earns. That applies to your positioning and to the automation you run: outbound that reads as machine-generated now lands with people who have already decided how they feel about machine-generated. Lead with the outcome and leave the mechanism in the footnotes.