Google Tests Conversational Ads That Actually Feel Like Conversations
Let's get something out of the way upfront.
We've been talking about "conversational ads" for years. Mostly, it's been a marketing buzzword slapped onto native placements that still feel like billboards with extra steps. Google's new Gemini-powered ad formats, unveiled at I/O 2026, are different. They're designed to weave directly into AI-generated search responses—both in Search and AI Mode—rather than bolt onto them.
The announcement landed quietly but matters enormously. Two new ad formats: Conversational Discovery ads and Highlighted Answers. Both use Gemini to make advertisements feel contextual, helpful, and genuinely conversational. Neither interrupts the user's flow. Both are currently in testing.
Here's what we know so far, and why this shift deserves a closer look.
The Two New Formats
Google's press materials at I/O 2026 didn't mince words: these are Gemini-powered ad formats built for conversational surfaces. That's the headline. The details are where things get interesting.
Conversational Discovery ads appear inside AI Mode when the system generates a response to a user's query. Instead of a sidebar or a floating banner, the ad is woven directly into the AI's answer text. Think of it as a sponsored suggestion embedded in a conversational response—Google's own product description frames it as closer to a recommendation from a knowledgeable friend than a traditional ad placement.
Highlighted Answers work similarly but target the standard Search results page. When Gemini generates an answer block, a sponsored response can appear as part of that block. Again, the key distinction: the ad isn't separate from the answer. It is the answer, with a sponsorship label.
Both formats rely on Gemini's language model to match ad content to user intent in real time. The ad copy isn't static. It adapts to the query, the context, and the conversational thread. That's a departure from every ad format we've used in search to date.
Why This Matters for Search Strategy
I'll be blunt: this is a paradigm shift, not an incremental update.
For decades, search advertising has been built on a simple transactional model. You bid on a keyword. You get a placement. You optimize your creative. You track conversions. It's linear. It's predictable. It's also becoming increasingly irrelevant.
Conversational Discovery ads and Highlighted Answers break that model entirely. Here's why:
Ads are no longer separate from answers. In traditional search, you know the difference between a sponsored result and an organic one. In AI Mode, that distinction blurs—or disappears entirely. The ad is embedded in the AI's response. The user isn't choosing between an ad and an answer. They're getting one thing that happens to be sponsored.
Static creative is obsolete. These formats use Gemini to generate ad content on the fly, tailored to the specific query and conversational context. Your ad copy doesn't get a fixed slot. It gets adapted in real time. That means your brand messaging needs to be flexible, modular, and designed for AI synthesis—not just human reading.
Placement is contextual, not positional. In traditional search, you fight for position zero, position one, or whatever the latest SERP feature du jour is. These new formats don't care about position. They care about relevance. If your content or product doesn't match the user's intent within the AI's generated response, your ad doesn't appear—regardless of your bid.
What This Means for Advertisers
If you're a paid search manager, your immediate reaction might be defensive. "How do I optimize for this?" "What's the bidding model?" "How do I measure success?"
Fair questions. The short answer: we don't have all the answers yet. Google is still testing these formats. But here's what I'd suggest teams start doing now:
Prepare your content for AI synthesis. If your ad copy is going to be adapted by Gemini in real time, your source material needs to be structured, clear, and AI-friendly. That means clean product data, well-organized brand messaging, and content that can be easily parsed and recombined by a language model.
Think in terms of relevance, not position. Traditional PPC strategy is obsessed with ranking. These formats reward relevance. If your product or service genuinely matches the user's intent within a conversational context, you have a shot. If not, no amount of bidding optimization will help.
Monitor AI Mode separately. A growing share of searches happen inside AI Mode. As of 2026, Google's AI Mode has billions of monthly users and is doubling every quarter. The conversational ad formats are being tested there first. If you're not tracking AI Mode performance as a distinct channel, you're missing a significant slice of the market.
Experiment with flexible creative. Static ad copy is the enemy of these formats. Start building modular content assets—product descriptions, value propositions, brand messaging—that can be recombined and adapted by AI without losing coherence.
The Bigger Picture
What Google is testing here goes beyond ad formats. It's a statement about the future of search.
The old model—keywords, bids, placements, conversions—is being replaced by a model where ads are generated contextually, woven into AI responses, and measured by relevance rather than position. It's slower. It's messier. It's also, frankly, more useful for users.
I've been skeptical of conversational ads from the start. Most of them feel forced, awkward, or like they're trying too hard to be "natural." Google's new formats are different because they're built on Gemini's language model from the ground up. The ads are the AI's response. They don't pretend to be anything else.
That's a big deal.
For advertisers, it means rethinking everything from creative strategy to measurement. For users, it means ads that feel less like interruptions and more like helpful suggestions. For Google, it means keeping users inside its ecosystem longer, with more useful, more contextual advertising.
The testing phase is just starting. But the direction is clear: ads are becoming part of the conversation, not separate from it. The question for advertisers isn't whether to adapt. It's how fast.
What to Watch Next
Google hasn't provided a rollout timeline for these formats. They're in testing. But based on Google's typical cadence, expect broader availability within 12–18 months. A few things to track:
- Bidding model: Will these use CPC, CPM, or a new pricing structure?
- Measurement: How will conversions be attributed when ads are embedded in AI responses?
- Eligibility: Which advertisers qualify? Will there be quality thresholds?
- User response: Will users engage with conversational ads, or will they feel intrusive?
Until we have answers to those questions, the best strategy is preparation. Build flexible content. Monitor AI Mode performance. Keep your product data clean and structured. And stay ready to adapt when Google opens the floodgates.
The future of search advertising isn't about buying placements anymore. It's about being relevant enough to be included in the conversation. That's a shift worth taking seriously.
For broader context on how Google's AI ecosystem is reshaping search strategy, see our analysis of agentic commerce and Google Ads and our breakdown of the new search landscape.
