A court filing confirmed what publishers feared — AI users won’t click links — yet Semrush data shows the visitors who do click are worth 4.4 times more than organic ones. Meanwhile, Meta launched an enterprise AI platform and Anthropic filed for an IPO that revealed frontier AI is still losing billions.
Court Filing Confirms: AI Users Won’t Click Your Links
An OpenAI engineer’s court brief cited Microsoft data showing Bing Chat delivers lower click-through rates than traditional search, openly admitting that AI answers displace the referral clicks that publishers and affiliate sites depend on. This is no longer speculation — it’s now on the record in a legal filing, which makes it hard for platform advocates to walk back.
Audit your top referral pages now and build direct audience capture — email lists, communities, bookmarks — before AI answers absorb more of your top-of-funnel traffic.
AI Search Visitors Convert at 4.4x the Value of Organic
Semrush’s latest SEO analysis found that visitors arriving via AI search convert at 4.4 times the value of a standard organic visitor, reframing the traffic-loss panic into a conversion-quality opportunity. Fewer clicks can still mean more revenue if the people who arrive are further down the purchase funnel and primed to act.
Stop measuring AI search purely by click volume — instrument your analytics to separate AI-referred sessions and track their conversion rate and revenue per visit.
CallRail Makes ChatGPT Ads Measurable for SMBs
CallRail launched a direct integration attributing calls, texts, and form fills to ChatGPT ad campaigns, then feeds that conversion data back to OpenAI to close the measurement gap that has kept smaller advertisers out of the channel. Without call-level attribution, proving ROI against paid search or social was impossible — this removes that blocker.
If you manage SMB or agency accounts, pilot CallRail’s ChatGPT integration on one campaign this quarter to establish a conversion baseline before the channel gets more competitive.
GPT-6 Astra Completes a 50-Tab Tax Workbook Twice as Fast
Tax software company Basis reported GPT-6 Astra finished a complex 50-tab tax workbook in half the time GPT-5.6 Sol required, with stronger intent understanding giving the team enough confidence to deploy it on live client work. This is a concrete benchmark on an error-sensitive, multi-step document task — exactly the category most teams have been reluctant to hand to a model.
Identify one complex, multi-step document workflow in your business and run a timed GPT-6 Astra pilot against your current process to get your own before-and-after number.
Meta Launches Enterprise AI Platform to Compete with OpenAI
Meta launched Meta Enterprise Platform — bundling Muse agent, Meta Business Agent, and Muse API and Code — and hired former MongoDB CEO Chirantan Desai as Chief Enterprise Platform Officer reporting directly to Zuckerberg. This moves Meta from ads-platform-with-a-chatbot into a direct competitor for enterprise AI budgets currently going to OpenAI, Microsoft, and Google.
If you already run paid campaigns inside Meta’s ad tools, request early access to Meta Business Agent to see how tightly it integrates with your existing account structure before competitors lock in an advantage.
Read the full story →
Join the discussion →
Anthropic Ships Sonnet 5.5 — and Says Retest Your Prompts
Anthropic released Claude Sonnet 5.5 alongside prompting guidance for Opus 5.5 explicitly warning developers to retest effort settings and reconsider instructions like “think carefully,” which behave differently on the new model generation. Prompts tuned for the previous generation can silently underperform or waste tokens — and Anthropic is saying so plainly rather than leaving teams to discover it in production.
Run your highest-traffic Claude prompts through a regression test on Sonnet 5.5 before pushing to production — treat each major model release as a breaking change until you verify otherwise.
Read the full story →
Read the full story →
Anthropic’s IPO Reveals $8 Billion Loss Behind Claude’s Pricing
Anthropic’s IPO prospectus disclosed $4.59 billion in revenue against an $8.06 billion operating loss, with $7.33 billion spent on compute and further large infrastructure commitments ahead. Today’s Claude API prices are subsidized well below actual cost, which means they are not a stable baseline for any multi-year budget or business model built on them.
Model your AI tooling costs with a 2–3× price increase scenario baked in, so a future repricing event doesn’t break a product or campaign you’ve built around current rates.
Hoka Owns the NYC Marathon Before Race Day Even Arrives
Hoka is running a sponsored Strava challenge tracking runners across New York City’s boroughs ahead of the November 2 marathon, using digital billboards as live scoreboards to extend brand contact across the entire training period — not just race day. Competing against Nike, New Balance, Brooks, and On for the running revival audience, Hoka built a multi-week owned engagement layer on a third-party fitness app where runners already spend time.
Look for the third-party platform your audience already uses obsessively and design a challenge or milestone mechanic that keeps your brand visible across weeks, not just a single campaign moment.
More from Rafal Reyzer
For deeper dives on AI and marketing strategy, visit my YouTube channel →
Rafal Reyzer
Hey there, welcome to my blog! I'm a full-time entrepreneur building two companies, a digital marketer, and a content creator with 10+ years of experience. I started RafalReyzer.com to provide you with great tools and strategies you can use to become a proficient digital marketer and achieve freedom through online creativity. My site is a one-stop shop for digital marketers, and content enthusiasts who want to be independent, earn more money, and create beautiful things. Explore my journey here, and don't forget to get in touch if you need help with digital marketing.