Ads Are Coming to Google AI Mode: What Advertisers Should Know

What is actually confirmed

On June 1, 2026, trade press reported that Google had begun a limited test of healthcare ads inside AI Mode. This is an early experiment, not a full ad launch, and the scope is narrow. But the signal is the story.

Two things are confirmed, and it is worth keeping them strictly apart.

First, the test. On June 1, 2026, Search Engine Roundtable reported that Google had begun a limited test of healthcare ads within AI Mode. Treat every word of that sentence as load-bearing. It is a limited test, in a single vertical, reported by trade press rather than announced as a general-availability product. Google has not, as of this writing, published ad formats, auction mechanics, pricing, or a rollout timeline for ads in AI Mode. Anyone telling you the format is "live" or describing exactly how the auction works is getting ahead of the evidence.

Second, the surface. At Google I/O 2026 (May 19-20, 2026), Google made AI Mode the default Search experience. Per Google's own framing, the company is upgrading Search with Gemini 3.5 Flash as the new default model in AI Mode for everyone globally, and AI Mode has surpassed one billion monthly users with queries more than doubling every quarter since launch.

AI Mode has surpassed one billion monthly users, with queries more than doubling every quarter since launch.Google - Search at I/O 2026

A surface that large, sitting at the front of Search, does not stay unmonetized. That is the context that makes a healthcare-ads test newsworthy rather than a footnote. Put those together and the honest read is this: ads in AI answers are an emerging reality, the healthcare test is an early data point in how they might appear, and the specifics are still unknown. For the broader picture on what the default-AI-Mode shift means for visibility, see our companion piece on Google AI Mode becoming the default search experience.

Why this matters even though it is just a test

A limited test in one vertical is still strategically significant for three reasons.

• The surface is now the default. AI Mode is not an opt-in tab anymore. With more than 1 billion monthly users and queries doubling quarterly, whatever ad model lands here eventually touches a large share of Google search demand. Decisions made in a healthcare test inform the template for everything that follows.
• Healthcare is a deliberate choice. Health is a high-stakes, heavily regulated, high-commercial-intent category. Google testing ads there first suggests it is solving for trust, disclosure, and compliance up front, which is exactly what it would need to do before expanding to other regulated verticals like finance, legal, and insurance.
• It changes the unit of the ad. A traditional search ad is a link plus copy beside a list of links. An ad inside a conversational answer sits next to, or inside, a synthesized response. The job of the ad shifts from winning the click against nine other links to being the relevant, trustworthy option the answer surfaces. That is a different optimization problem, and the teams who start thinking about it now will adapt faster.

How ads in AI answers might work

With that label firmly attached, here is the most reasonable planning hypothesis.

• Disclosure will be prominent. Google's long-standing practice is to label paid placements clearly, and it has signaled the same intent for ads in AI experiences. Expect a visible "Sponsored" treatment, likely more conspicuous in a high-trust context like health.
• Relevance and quality signals will gate eligibility. Ad rank in conventional search already blends bid with quality and expected impact. It is reasonable to expect that ads surfaced within an AI answer will lean even harder on relevance, landing-page quality, and policy compliance, because an irrelevant ad inside a synthesized answer is more jarring than one in a list.
• The ad may be tied to the answer's intent, not just the keyword. Conversational queries are longer and more contextual. Targeting may key off the resolved intent of a multi-turn exchange rather than a single matched keyword, which would make broad-but-controlled match types and strong signals more important than exact-match keyword lists.
• Feeds and structured product or landing data will likely do more work. When an answer is assembled rather than listed, the engine needs clean, machine-readable data about what you offer to decide whether and how to surface you. This is the same direction Google has pushed for Shopping and agentic commerce.

If Google publishes formal documentation, it will appear in the Google Ads Help Center. Watch it, and replace these assumptions with facts as soon as they exist.

How advertisers should prepare now

You cannot optimize a format that has not shipped. You can put your account, data, and measurement in a state where you adapt in days instead of quarters. Here is the preparation work that pays off regardless of the final format.

1. Get your product, landing, and feed data clean

If AI answers are assembled from structured data, the quality of your machine-readable inputs becomes a ranking and eligibility input, not just a hygiene nice-to-have.

• Audit your Merchant Center and product feed for completeness, accuracy, and freshness. Fill optional attributes, not just required ones.
• Make sure landing pages match ad and feed claims exactly. Price, availability, and offer parity matter more when an AI summarizes them.
• Add and validate relevant structured data on key pages. For commerce, that means complete Product and Offer markup; for services, accurate Service and Organization data. Clean structured data is how machines understand what you sell.

2. Strengthen landing pages for qualification, not just clicks

Placement inside a conversational answer likely means fewer, more-qualified visitors who arrive with more context. Optimize for that.

• Make the page deliver immediately on the specific intent the answer addressed. Vague, broad landing pages will convert worse when the visitor was just handed a precise answer.
• Reduce friction: fast load, clear next step, trust signals visible above the fold.
• Treat the page as the place where the blurred click gets re-earned. If users research inside the AI answer and arrive ready, the page's job is to confirm and convert, not to re-pitch from scratch.

3. Double down on measurement and incrementality

This is the most important preparation item. As AI answers absorb research and comparison steps, the path from impression to conversion gets harder to see. Last-click attribution will understate the value of upper-funnel exposure inside answers.

• Move toward data-driven attribution and, where you can run it, geo-based or holdout incrementality testing. Incrementality tells you what the spend actually caused, which matters when the click is blurred.
• Make sure conversion tracking is robust: server-side tagging, enhanced conversions, and clean offline-conversion imports so qualified leads and sales tie back to spend.
• Define success in business outcomes such as qualified leads, revenue, and CAC rather than clicks or CTR, which will be less interpretable in an answer-first surface.

4. Keep account hygiene and a testing posture ready

When a new format opens, the accounts that already have clean structure, good signals, and a culture of disciplined testing capture early-mover advantage.

1. Tighten account structureMaintain tight account structure, current negative-keyword lists, and accurate conversion goals so a new inventory type can be adopted without untangling old debt.
2. Reserve a test budgetReserve a test budget and a measurement plan you can deploy quickly. Treat any AI-answer ad beta as an experiment with a clear hypothesis and a holdout, not a set-it-and-forget-it line item.
3. Watch policyNew surfaces ship with new policy, and regulated verticals ship with the strictest version of it. Read the policy the day a beta opens, not after a disapproval.

For context on why the click itself is changing, our analysis of zero-click search and publisher traffic in 2026 covers the same dynamic from the organic side. The measurement lesson applies to paid too.

Why watch the healthcare test specifically

The healthcare test is a bellwether. Health is a Your Money or Your Life category: Google holds it to a higher bar for accuracy and trust, and healthcare advertising already carries significant regulatory sensitivity around claims, eligibility, and data handling. If Google can ship ads into AI answers in a category this scrutinized, with disclosure and compliance that hold up, it has effectively built the playbook for rolling ads into other regulated and high-intent verticals.

So watch for the things that travel across verticals. None of these is confirmed for a broad rollout yet, but observed carefully in a test they are the best available preview of the general format.

• DisclosureHow the ad is labeled, and how prominent the "Sponsored" treatment is in a high-trust context.
• PlacementWhere the ad sits relative to the synthesized answer - beside it, inside it, or after it.
• TargetingWhether the ad keys off conversational intent or a matched keyword.
• ControlsWhat advertiser controls and reporting Google exposes for the new inventory.

For the full slate of I/O 2026 changes and how they connect, see our complete Google I/O 2026 breakdown for search marketers.