AI Search & GEO News

Parsnipp commentary

The revenue data here is the real headline: shoppers referred by LLMs already spend 53% more per visit than non-AI traffic, even while only 11% of consumers say they'd let AI buy fully on its own. That combination is exactly our thesis — invest now in being discoverable and accurately represented everywhere AI pulls from, because the research phase is already monetizing even before autonomous checkout catches up.

Parsnipp commentary

Worth noting Colgate-Palmolive's approach: this isn't a marketing-team side project, it's three years of data infrastructure work plus training 70% of VPs on AI before touching agentic commerce. It's a good reference point for how much unglamorous data-hygiene work usually sits underneath a brand's "AI readiness" — the same groundwork our GEO audits tend to surface for clients.

Parsnipp commentary

Now, OpenAI is letting its own systems decide single-product vs. carousel placement rather than leaving it to advertisers — another sign that ad formats inside AI answers are maturing fast, timed deliberately for Q4 budget planning. Brands sitting out "LLM ads" while they mature the format risk playing catch-up the way early paid-social laggards did.

Parsnipp commentary

The stat that should worry every brand: 68% of Google searches now end with no click-through, and ChatGPT's share of the AI answer market dropped from 78% to 56% in a year as Gemini and Claude grew. Visibility now has to be measured and defended across multiple AI engines at once — optimizing for just one model is already out of date by the time you finish.

Parsnipp commentary

Good tension named here: AI needs clean, structured product data, but consumers still buy on meaning and aspiration — brands that only optimize for the algorithm and forget the human on the other end will win visibility but lose the sale. This is the balance we push clients toward: fix the data plumbing, but don't let it replace the brand story.

Parsnipp commentary

AI Overviews jumped from 15% to 43% of Google searches in a single year — that's not a slow transition, it's already the default experience for a huge share of queries. If your GEO strategy is still "wait and see," this is the data point that ends that debate.

Parsnipp commentary

Google is moving from "answer once" to "keep watching and tell me when something changes" — search that stays alive after the first query and follows you across devices. That changes the optimization target again: brands need to think about being the answer that surfaces later, automatically, not just the one that ranks at query time.

Parsnipp commentary

Two ads from two different advertisers in one response footer is the ad-inventory equivalent of "getting crowded" — a signal ChatGPT's ad demand is scaling fast, not just testing. Brands waiting for LLM ads to feel more "proven" before jumping in are watching the window to be an early mover close.

Parsnipp commentary

This is the plumbing layer most brand marketers aren't thinking about yet: if AI agents are going to transact for people, they need agent-native money rails, not just agent-native discovery. Worth keeping an eye on, since payment friction (or lack of it) will shape how fast "agent buys it for you" actually scales.

Parsnipp commentary

Another data point that agent-to-agent commerce is being built out as real infrastructure, not hype — VCs are funding the "how do agents actually pay" problem directly. The more this layer matures, the faster the gap closes between an AI agent recommending your product and one buying it outright.

Parsnipp commentary

Meta's shaky AI ad creative and the shift of human media buyers into "quality-control" roles both point to the same lesson — AI automation still needs a human checking its work. Same logic applies to AI search visibility: automate the monitoring, but don't stop reviewing what it surfaces.