AI search gets measurement teeth; ad tech braces for remedies

Add the Lighthouse ARD audit to your technical QA checklist now, but track the spec diffs — teams that hardcode against 13.5's checks will rework when the proposal settles.

If AI Mode impressions are climbing in your Search Console data, rebuild your click-through assumptions by query class now — the old organic CTR curves overstate remaining traffic for AI-served queries.

Inventory which AI crawlers you currently allow versus block before letting a category-level default decide for you — separate training access from search access or you'll accidentally grant both.

Sit out the pilots if you can, but document your content's licensing posture internally now — whichever model wins, publishers with clean usage data and defined terms will negotiate from strength.

Audit canonical tags across every domain you control, including staging and syndication partners — a stray cross-domain canonical is the fastest way to silently de-index a revenue page.

Don't rip out Google stack integrations yet — the remedies phase in slowly. But add header-bidding and alternate-exchange yield tests to your Q4 roadmap so you have data when interoperability actually bites.

Visual Search Ads are worth a pilot if you have strong product imagery and apparel or home SKUs; format fit matters more than budget here, so start with your best-photographed category.

If Kokai migration stalled for your accounts, press your TTD rep on timelines — a 15% workforce cut rarely accelerates product roadmaps, and your media plan shouldn't assume it will.

If you buy location-targeted media through either platform, get contract continuity terms in writing before the integration — data-migration handoffs are where targeting quality typically drops.

Watch which publishers sign direct AI feeds — if your content syndication partners go this route, your citations may start flowing from licensed feeds you don't control. Map who owns your distributed content.