Adrian Rosenkranz’s core message was that go-to-market scaling now depends on understanding how AI is changing discovery, traffic, and content creation in near real time. Speaking as Chief Revenue Officer at Webflow, he framed the session around what Webflow is seeing across more than 400,000 customer websites and how those patterns are forcing revenue leaders to become more detail-oriented and adaptive. He explained that when he joined Webflow roughly eighteen months to two years earlier, he expected to apply familiar playbooks from Tableau and Salesforce around outbound, team selling, and enterprise expansion. Instead, AI changed the environment: large language models began searching the web, customer research behavior started shifting monthly, and the path by which buyers find and evaluate companies became less predictable.

He organized the change into three buckets: bots reading before people do, zero-click marketing accelerating, and AI changing how teams build content. On bot traffic, he shared that Webflow has seen an almost 62% increase since January across its customer base, with the acceleration especially visible in the summer months. He attributed much of that summer increase to Meta aggressively crawling after releasing Muse, alongside increased crawling from ByteDance. He then contrasted this with human traffic patterns: direct and unknown traffic rose significantly, while social and search traffic fell. His hypothesis was that more buyer research is happening outside company websites, while the website still acts as a source of truth that informs social platforms, search engines, and AI assistants.

The zero-click discussion centered on declining click-through behavior. Rosenkranz said that in January, social and search might drive three to four human visits for every bot visit, but the ratio is moving closer to one human click for each bot crawl. He linked this to behaviors such as people consuming previews on LinkedIn or Instagram without visiting the underlying page. AI referrals also changed: from January to August they rose by two percentage points overall, but the more important pattern was that they climbed close to 8% before falling. He highlighted a shift between AI models actively searching the web and models already having the answer, arguing that this means companies must keep their website content current because they cannot know when an AI system will next check it.

In the final provided portion, he turned to AI-assisted content creation. Webflow saw MCP adoption, defined as using AI to talk to Webflow, increase 2.7 times since April, and also saw faster publishing when AI was used as part of creating website content. However, he warned that the environment is changing again: three months earlier there was little penalty for creating with AI, but platforms and audiences are beginning to flag low-quality AI-generated material, such as LinkedIn’s “AI slop” signals. An audience question clarified that the data came from Webflow’s global base of 400,000 websites, spanning customers from individuals and scale-ups to companies such as Amazon, Nvidia, Reddit, monday.com, and Claude.com.