Google's AI Search adds links, Ads adds journey bidding, ChatGPT ads go global

Treat AI Search link presence as a brand-visibility metric, not a traffic metric. Build a citation-tracking workflow off log files and prompt audits, because Search Console will not show this surface any time soon.

If your media plan touches any of these five markets, get on the test list now. Pricing and measurement norms will be set by the first cohort of advertisers, not by whoever shows up in 2027.

Map your high-value pages against Bing's five grounding areas, especially freshness and verification. Sites that pass classic SEO audits but lack timestamped, source-cited content tend to lose grounding citations even when they rank.

Before adding a GEO line item to a 2026 budget, ask the vendor which specific behaviors in LLM citations their tactic moves. If the answer is generic, the work belongs in your existing SEO retainer.

Stop pitching Preferred Sources as a brand-defense play to clients. It is a personalization feature, not a ranking shortcut, and quality remediation work still has to come first.

Audit your product feed coverage and structured data for items where the SERP transaction will replace a site visit. Last-click attribution will understate Google's contribution, so add a UCP-aware view to your weekly ecommerce report.

Re-baseline your tCPA and tROAS targets before turning Journey-aware Bidding on. Historical conversion paths will be reweighted, and accounts with thin conversion volume will see the most variance in the first two to four weeks.

Ask your SSP and DSP partners for their actual agentic campaign counts and outcomes, not their roadmap decks. The gap between PubMatic's 30 live campaigns and most vendor pitches is where the next 18 months of programmatic budget will be decided.

If you have been holding off on AI Max because of brand safety on query expansion, the new steering controls are worth a structured pilot. Run it on one ad group with negative keyword and brand-list guardrails before expanding.

Add an AI-influence row to your monthly client reporting now, even if the methodology is rough. The teams that wait for clean attribution will lose budget arguments to teams who can show directional data this quarter.

OpenAI is staffing for a real ad business, not an experiment. Brief your CMO that ChatGPT inventory will likely have credible measurement and identity stories within 12 months, and plan test budgets accordingly.

Regulated and government clients should track AP2 governance now, not after it lands in browsers. Procurement, accessibility, and identity teams will all have a stake once agent-initiated transactions hit production sites.