This week: Google rolled out AI Mode in India without requiring Labs signup, while also releasing enhanced forecasting tools in Keyword Planner; TikTok is developing a separate US version of its app ahead of a possible forced divestiture; Cloudflare introduced Pay Per Crawl to let publishers charge AI crawlers; and a new study revealed that AI tools get PPC advice wrong 20% of the time.
Google Keyword Planner now offers enhanced forecasting tools that break down keyword forecast volumes by city, region, and device type, bringing a new level of granularity to campaign planning.
For advertisers targeting multiple locations or optimizing for specific user contexts, this update enables more precise budgeting and local strategy alignment. The ability to forecast at the municipality level can help you identify high-performing geographic areas before launching campaigns.
TikTok is reportedly developing a separate version of its app for US users, with a planned launch date of September 5, while the current app would stop functioning in the US by March 2026.
Rollout of a US-specific app could impact audience targeting, ad delivery, and data continuity. A forced app transition might disrupt user behavior, shrink available data signals, or reset algorithmic learning, requiring marketers to adjust campaign strategies and prepare for potential platform fragmentation.
Cloudflare launched Pay Per Crawl, a first-of-its-kind monetization layer that lets publishers charge AI bots (like OpenAI's GPTBot or Anthropic's ClaudeBot) for accessing their content, with major publishers including TIME, The Atlantic, and BuzzFeed already joining.
This new system could fundamentally change how AI models access training data, potentially impacting how brands appear in AI-generated responses. For advertisers, this means the content you create may need to be more explicitly designed to break through the wall garden if publishers widely adopt this approach.
A Wordstream study testing five popular AI tools (ChatGPT, Google AI Overviews, Google Gemini, Perplexity, Meta AI) found that one in five PPC-related answers were flat-out wrong, with Google Gemini being the most accurate at 94%.
While widely used in PPC, AI tools are often wrong enough to be concerning – representing potential revenue loss for advertisers. As AI becomes more embedded in marketing workflows, knowing which tools to trust (like Gemini over others) and when to verify answers is essential for protecting ad spend.
Google Ads' ad sharing, which lets a single ad be reused across multiple ad groups, will be deprecated with v22 of the API, with creation of shared ads blocked from mid-October and existing shared ads phased out entirely next year.
This is a major shift for developers and advertisers managing campaigns at scale through the API. The removal of ad sharing means advertisers can no longer reuse a single ad across multiple ad groups, which impacts workflow efficiency and requires updating systems to support Google's asset-based ad formats.