Follow the Spend:

the NBA Finals were Google's biggest AI announcement

If you followed the NBA Playoffs this year, you saw it. Google was everywhere. Player partners like Luka Dončić, Stephen Curry, and Shai Gilgeous-Alexander breaking down their game through AI Mode. A sustained run of live, in-game integrations across NBC, Disney, and Prime, nudging millions of viewers to ask Google's AI instead of searching the old way. The whole run built to the Finals, the most-watched window of the basketball year.

When a company spends like that around its biggest sports partnership of the year, it is not a branding exercise; it is a signal about where Search is headed.

Here is the backdrop: AI Mode, Google's conversational, answer-first version of Search, just crossed 1 billion monthly users. At Google Marketing Live in May, Google announced it is putting ads directly inside that experience: AI-powered Shopping ads, Conversational Discovery ads, and a Business Agent that turns a static lead form into a live chat. The search interface brands have optimized against for two decades is becoming a conversation, and Google is bringing the ad inventory along with it.

Why this matters for your business

Roughly two-thirds of searches now end without a click, and in AI Mode specifically, that figure climbs past 90%. When the AI answers the question on the page, the top-of-funnel and research-stage traffic you used to capture never reaches your site. That traffic is not disappearing evenly. It is collapsing into the results page itself.

The knock-on effect shows up in your auction. As exploratory clicks dry up, paid competition concentrates on the high-intent, revenue-dense keywords that still convert. More bidders chasing fewer valuable clicks means more CPC volatility, especially in margin-sensitive categories. If your costs have felt jumpy lately, this is part of why.™

What we are doing about it

Feed and content health is now a ranking factor in disguise. Gemini writes its own explanation for why your product fits a shopper's question, and it pulls that from your feed and your site. Thin content means you do not get surfaced. This is exactly what our generative engine optimization (GEO) framework is built for: disciplined feed management plus content that gives the AI strong signals to pull from. It is the highest-return work available right now.

The high-intent keywords that still drive revenue need tighter management, not autopilot. That is where the volatility lives, and where a few points of efficiency compound fast.

We are also preparing for the new AI Mode formats now. They are rolling out in stages. We run a structured testing framework for new inventory sources, so when a placement opens, we move in early with clean data and creative instead of scrambling after the fact. Getting into a new placement before the auction gets crowded is one of the few remaining structural advantages in paid search. Google is building this in real time, and so is every buyer. What sets our position apart is that we are not starting from theory. We have a growing body of learnings from live client data across verticals, and the approach we bring to clients is grounded in what is actually moving performance, not what should move it on paper. The direction is set, and a Finals-sized ad budget removes any remaining doubt about how committed Google is.

The value of a search program is shifting away from keyword management and toward staying ahead of where traffic is going. That is the conversation worth having now, before the formats are fully baked and the early-mover advantage closes.