What actually matters from GML 2026, why it matters for your business, and the concrete moves marketers should make right now. AI Mode hit a billion users. Agentic commerce is real. Here's what to do about it.
Five themes from GML 2026 — and the move you should make on each one.
AI Mode hit 1B users. The query is now a conversation.
UCP changes how buyers move from intent to purchase.
Brand and performance on one platform. The math finally works.
Gemini + Asset Studio = fewer bottlenecks, better ads.
Your data is now competitive infrastructure.
I'm writing this as someone who has built demand-gen engines from scratch at B2B and B2C companies. GML 2026 isn't a product release — it's a signal about where the leverage is shifting. The marketers who move on AI Mode, Demand Gen, and first-party data infrastructure in the next 90 days will have compounding advantages that are very hard to close. Here's what I'd prioritise.
AI Mode passed 1 billion monthly users at GML 2026. That's not a beta product — that's the product. AI Mode queries run 3x longer than traditional searches, which means buyers are doing more research, surfacing more nuance, and expecting more specific answers before they act.
| Feature | What it does | Why it matters |
|---|---|---|
| AI Max for Search | Matches ads to longer, more nuanced queries automatically | Your existing keywords no longer define your reach ceiling |
| Performance Max | Unified campaign type across Search, Display, YouTube, Maps | One budget, AI-optimised across surfaces — less manual arbitration |
| Direct Offers | Special offers served inside AI Mode conversations | Captures buyers at their highest-intent moment |
| Agentic ads (chat) | Conversational ads built with Gemini, inside Search | Leads arrive pre-qualified; the ad does some of the selling |
Google's guidance: lead with what only your brand can say, focus on genuinely helpful content, and be agent-ready. That last point is new — it means your content needs to be structured so AI agents can parse and surface it in conversational contexts, not just ranked by a blue-link algorithm.
AI Mode is not a reason to abandon traditional SEO — it's a reason to raise your content quality floor. Thin, keyword-stuffed pages don't get cited in AI answers. Substantive, expert, specific content does. Audit your top 20 pages for 'what can only we say here?' and close the gap. That's the compounding advantage.
The Universal Commerce Protocol (UCP) was the most strategically significant announcement at GML 2026. It's a new industry standard — already signed on by Shopify, Etsy, Wayfair, Target, Walmart, Amazon, Meta, Microsoft, Salesforce, and Stripe — that lets AI agents complete purchases on behalf of buyers without custom integrations.
UCP is a common language that connects AI shopping agents directly to merchants' inventory, loyalty data, and checkout systems in real time — no custom code required. When a buyer asks an AI assistant to 'find me running shoes under $120 that ship before Friday,' the agent can search, compare, and transact without the buyer ever leaving the conversation.
UCP sits between the AI agent and the merchant's backend. It exposes live inventory, pricing, loyalty status, and checkout rails as a standardised data layer — so any UCP-compatible agent can transact with any UCP-compatible merchant. Merchants retain full control over their data and relationships.
Google's explicit guidance: rich Merchant Center feeds with complete product details are the foundational requirement for conversational discovery. A product not in your feed is a product that can't be recommended by an AI agent. Feed quality is no longer a technical nicety — it's a distribution lever.
If you run e-commerce or sell physical products, feed quality is now a revenue-critical infrastructure project — not a marketing ops task. Treat your Merchant Center feed the way you treat your CRM. And if you have 1,000+ SKUs with incomplete attributes, that gap is costing you reach in every AI-mediated shopping surface starting now.
YouTube reached 91% of the US adult population 18+ in a single month (November 2025, per Nielsen). That's 244 million people. GML's core YouTube message: you no longer have to choose between building brand equity and driving direct response — Demand Gen does both in the same campaign.
Adding Demand Gen to Search and Performance Max campaigns drives 10% higher ROAS and 12% higher sales effectiveness on average (Nielsen MMM meta-analysis commissioned by Google, 2024). In H2 2025, Google made hundreds of improvements to Demand Gen that drove a 30% increase in conversions on average. New GML additions include Maps inventory, expanded Product Feeds, and creator asset integration.
| Metric | What it measures | Why it wins CFO buy-in |
|---|---|---|
| Engaged View Conversions | Someone chose to watch your ad before converting | Proves intent, not accidental exposure |
| Campaign Type Attribution | Every conversion Demand Gen contributes to, cross-campaign | Apples-to-apples against Search and PMax |
| 86% long-term ROAS lift | Brand equity impact beyond the initial 30 days | Makes the case for sustained investment, not just ROAS in week 1 |
The 45% of YouTube Shorts users who aren't on TikTok, and the 65% not on Instagram Reels, are an audience you genuinely cannot reach anywhere else at scale. If your paid social is heavily Meta-weighted and you haven't seriously tested Demand Gen, that's the single highest-leverage experiment on your roadmap this quarter.
Creative is the largest driver of ad effectiveness — contributing to nearly half (49%) of incremental sales, per NCSolutions 2023. GML 2026 announced tools to close the gap between 'we know creative matters' and 'we can actually produce enough of it.' Asset Studio is the hub.
Asset Studio now integrates Gemini and Veo (Google's video AI) directly, alongside Product Studio, Pomelli (a new Google Labs tool for on-brand creative assets for small businesses), Canva, and Adobe. Everything connects to one production hub with built-in A/B testing and brand-guideline controls.
| Tool | What it produces | Best for |
|---|---|---|
| Asset Studio + Gemini | Text and image ad assets, RSAs, Display | Teams who need volume without growing headcount |
| Asset Studio + Veo | Professional-grade video assets in minutes | YouTube Demand Gen creative at scale |
| Product Studio | Product-focused creative, lifestyle imagery | E-commerce SKU-level creative variation |
| Pomelli (Google Labs) | On-brand assets for small businesses | Lean teams with limited design resources |
| Canva / Adobe | Third-party design feeds directly into Studio | Brands with existing design workflows |
Creative volume is no longer the constraint — creative quality and brand control are. The teams who win with these tools will be the ones who invest in setting up strong brand inputs (voice guidelines, approved imagery, messaging pillars) so the AI generates on-brand assets, not generic ones. Garbage in, garbage out still applies.
GML's measurement message was pointed: data strength is no longer an operations function, it's a competitive advantage. The better your first-party data, the better your AI campaigns perform. This compounds — and the gap between brands who take it seriously and those who don't will widen every quarter.
| Pillar | What Google launched | What to do |
|---|---|---|
| Data Strength | Data Manager + Google Tag Gateway for signal consolidation | Audit every data source feeding your campaigns; close the gaps in first-party signal before Q3 |
| Causality | Attributed Branded Searches (ABS) for short-term intent; Qualified Future Conversions (QFC) for long-term proof | Use ABS to show brand-building is working on a short cycle; use QFC to justify brand investment to finance |
| Unified View | Meridian (open-source MMM) now integrated into Google Analytics 360 as a data command centre | Run Meridian on your H1 data before the next planning cycle; stop relying on last-click attribution alone |
Dr. Martens fed first-party margin data directly into Performance Max to optimise campaigns for actual product profitability, not just conversion volume. The result: 16% revenue growth. The lesson — when you give the AI better signals, it optimises for better outcomes. Most marketers are still optimising for the easy conversions, not the profitable ones.
Most marketing teams underinvest in data infrastructure because it doesn't ship a campaign. But your data quality is now the ceiling on your AI performance. Every campaign you run on incomplete signals is leaving performance on the table. The one project I'd fund before any new campaign budget: close your first-party data gaps and deploy Google Tag Gateway.
If I were running a marketing team right now, these are the five moves I'd prioritise in the next 90 days — in order of leverage.
Review each for unique, expert content that an AI would actually cite. Flag and close the thin content first.
Complete every product attribute — title, description, image, GTIN, price, availability. A complete feed is now a discovery prerequisite.
Add Demand Gen to one existing Search or PMax campaign. Give it the 4-week minimum. Measure using Engaged View Conversions and Campaign Type Attribution, not last-click.
Write a one-page brand brief: voice, visual rules, approved messaging, what to avoid. This is what keeps AI-generated creative on-brand at scale.
Audit what's actually flowing into your campaigns vs what should be. Fix the highest-volume gaps first. If you're not running Meridian or an equivalent MMM, start the evaluation.
Build the engine, not the campaign. If you're a marketing leader trying to build the systems that compound, that's what I do.
hello@kerrycokelekoglu.com kerrycokelekoglu.com