This week, Google and OpenAI made moves that, at first glance, seem unrelated. One upgraded its AI model. The other launched business profiles. One added transparency to Google Ads. Meanwhile, the affiliate model is being squeezed by AI-powered checkout.

But line them up and they say the same thing: AI platforms are no longer experiments. They are infrastructure. What you do with them now determines whether you will still be visible in six months, or whether you will realize too late that the rules changed without you.

Here are the four developments we are watching at difrnt. this week.

ChatGPT Business Hub: your business profile, now in AI too

On November 10, OpenAI launched Businesses in ChatGPT in beta. Companies can now claim a profile directly inside ChatGPT, similar to Google Business Profile, with one key difference: it accepts online-only businesses with no physical location.

Verification requires adding a DNS TXT record to prove domain ownership, plus identity verification and business details (legal name, address, registration ID). OpenAI states that having a profile does not guarantee better placement in responses. But the data speaks for itself: Search Engine Journal reported one publisher registering 353,356 mentions and 2,663 website clicks in a single month through this system.

If Google Business Profile became essential for local SEO, ChatGPT Business Hub could become the equivalent for GEO. The signal is clear: OpenAI is building a business directory inside ChatGPT, and early claimants get first-mover advantage.

The cost is zero. Setup takes minutes. And this is not just about local businesses anymore. The risk of not being there when millions of users ask ChatGPT "what is the best marketing agency near me" is significantly higher than the effort to get listed.

Gemini 4 Hydrogen: the model that changes what you see in Google

On November 12, Google made Gemini 4 Hydrogen the default model in AI Mode for all users worldwide. The same model now handles the most complex questions in AI Overviews as well.

What makes Hydrogen different? It promises better answers to complex, multi-part queries, including those with typos. And a "query fan-out" that, according to Google, connects users to a broader range of helpful links. What that means in concrete link volume terms, Google has not specified. Hydrogen is built for speed and efficiency, unlike Gemini 4 Carbon and Argon, which handle deep reasoning tasks and are available only to paying subscribers.

A strategically important detail: the switch came just two days into the November Core Update. Any citation data from AI Mode or AI Overviews collected this month spans a model change in the middle of a major algorithm update. If you track AI Search visibility, treat November data with extra caution.

The bigger message: Google is accelerating AI integration in search. Hydrogen is fast and efficient, built for the masses, not the niche. More users get AI-generated answers, fewer click through to traditional results. If your brand does not appear in those answers, the problem is not the algorithm. It is your content strategy.

Google Ads now shows invalid traffic. Partially.

Google Ads is rolling out an Invalid Traffic Overview Card on the Overview page, available globally and in Search Ads 360. The card shows estimated savings from invalid traffic filtered before billing, post-billing credits, and campaign-level reports downloadable as CSV.

An important note: post-billing credits do not change the Cost or CPA in the main Campaigns table. For adjusted costs, you need to access the "Invalid Activity Credit: Search & PMax" report in Report editor. An extra step that many advertisers will overlook.

Why does it matter? Not because it solves the invalid traffic problem, but because it provides a level of transparency advertisers have not had before at this level of detail. Until now, you could see your costs but not how much was filtered invalid traffic. It is a small step toward transparency, but the first of its kind at this level of granularity.

But do not get comfortable: the card does not show undetected invalid traffic. Higher-spending campaigns may show larger savings simply because they have more total traffic, not because they are "cleaner." Lead quality still needs verification in your CRM, not just in the Google Ads dashboard. Add this card to your monthly campaign review as an additional data point, not as a reason to relax.

Does the affiliate model survive AI checkout?

Google's UCP now enables checkout directly inside AI Mode and Gemini. TikTok Shop and ChatGPT's retailer apps keep purchases entirely on-platform. What happens to affiliate links when the user never reaches your website?

The link remains the first line of defense for referral credit. But tracking has evolved well beyond cookies: cross-device syncing, server-to-server tracking, IP hashing, vanity codes. Each comes with its own limitations. Vanity codes, for instance, leak quickly to coupon sites and dilute attribution. The entire attribution model that affiliate marketing was built on is being challenged by platforms that want to keep users inside their own ecosystem.

An interesting element: in a world where consumers distrust AI recommendations, human-made reviews are becoming more valuable than ever. Affiliate disclosure is no longer just a legal obligation but an authenticity signal. "Your click supports a real creator" is a message that works in 2026.

If you work with affiliates or are one, prioritize programs that pay commissions on marketplace and third-party sales, not just direct ones. Monthly reconciliation of unattributed sales has become standard practice, not an exception. In a market where organic traffic is declining, diversifying revenue sources is not optional.

Four moves, one message

AI is no longer a marketing channel you can ignore while relying on classic SEO and Google Ads. It is infrastructure that redefines how you are found, evaluated, and chosen. Google and OpenAI are making weekly moves that change rules we considered stable. And neither company is waiting for the industry to catch up.

The question is not whether to invest in AI presence. It is why you have not started already.