Google’s Gemini-powered Conversational Discovery ads reward deep technical knowledge over generic brand claims. Here is how B2B businesses can win in AI-driven search.
Google has started testing Conversational Discovery ads, a Gemini-powered format that answers a user’s specific question inside AI-driven search experiences. For B2B businesses with genuine technical depth, this is quietly the best news to come out of Google Ads in years. The old game of who could shout the loudest with the biggest brand budget is being replaced by something more interesting: who can actually answer the question in front of them.
I have spent a lot of time watching ad formats come and go, and most of them ask you to strip your knowledge down to a headline and a call to action. This one does the opposite. It rewards the businesses that know their subject inside out, because the AI has to draw on that expertise to build a useful answer. If you have ever felt that your deep product knowledge was wasted on a 30-character headline, this format is worth understanding properly.
What exactly are Conversational Discovery ads?
Conversational Discovery ads are a new advertising format that Google began testing in 2026. Rather than serving a static headline and description, the ad responds to a person’s specific question inside an AI-driven search experience, using Gemini to build creative tailored to that individual search. In plainer terms, when someone asks a detailed question, the ad can form part of a genuine answer rather than sitting beside the results hoping to be noticed.
This sits alongside the broader Gemini-powered conversational experience that Google has been rolling out inside Google Ads itself. That experience is described by Google as a chat-based feature designed to combine your expertise with Google AI, and it initially expanded to English-language advertisers in the U.S. and U.K. before a wider release. You can read Google’s own framing of how these AI-native formats are meant to help brands connect with consumers in the Google Marketing Live search ads announcement.
The important word in Google’s description is ‘combine’. The AI is not replacing your input, it is amplifying it. If the underlying expertise you feed the system is thin, the answer it generates will be thin too. If it is rich, specific and grounded in real product knowledge, the format has something worth working with.
Why does technical depth beat generic marketing in Gemini-era ads?
For years, a lot of B2B advertising leaned on brand signals: awards, sponsorships, recognisable logos, confident but vague claims about being a market leader. That approach was always weaker than it looked, and the shift towards conversational, question-led ads exposes the weakness more clearly than ever.
A 2024 manufacturing buyer journey study put it bluntly: technical expertise trumps brand. Buyers prioritised demonstrable technical and industry-specific expertise over general brand recognition, awards or sponsorships. The same study found that a third of buyers felt vendors often lacked sufficient technical depth and clear product differentiation. That gap is precisely the space a well-informed conversational ad can fill.
Think about what happens when someone asks a genuinely specific question, the kind of query a technical buyer actually types. A generic brand message has nothing useful to say. A response built on real knowledge of tolerances, integration requirements, implementation realities or compliance constraints can address the question directly. Google’s support documentation notes that the conversational experience helps create tightly themed ad groups and optimised responsive search ads, which is direct evidence that your domain knowledge still shapes performance even as the AI handles more of the generation. The businesses that will struggle are the ones whose marketing was always a layer of gloss over a shallow understanding of their own product.
What do B2B buyers actually expect before they contact you?
The reason this format matters is that buyer behaviour has already moved. People are doing far more of their research before they ever speak to a salesperson, and they expect to find real substance while they do it.
A 2025 Responsive report found that 90% of buyers conduct research before first contact, and nearly two-thirds are already using generative AI tools as much as, or more than, traditional search. That is a fundamental change in where the buying journey begins. If most of the evaluation happens before anyone reaches your website, the quality of the information you surface in those AI-mediated moments becomes the thing that gets you shortlisted. You can see how buyers describe this shift in the Responsive report on the B2B buyer’s mind.
Buyers are also increasingly demanding when it comes to self-service. A 2025 Contentful benchmark found that 84% of B2B buyers say self-service tools are critical when choosing a vendor, and their top frustrations included a lack of pricing transparency, time-consuming reordering and slow vendor response. All of that points in the same direction: detailed, practical, honest information beats polished vagueness. And when buyers do compare options, integration tends to matter more than standalone features. A 2025 Demandbase report found buyers prioritise integration capabilities over standalone features, which supports a technical, implementation-focused message rather than a feature checklist.
What kind of information should you be feeding the system?
Practically, this means documenting the things your team knows instinctively but rarely writes down. How your product integrates with common systems buyers already use. What implementation actually involves, including timescales and the awkward bits. Where your offering genuinely fits and, just as usefully, where it does not. The specific problems you solve and the language buyers use to describe them. This is the raw material a conversational ad needs to answer a real question well, and it is exactly the material most businesses keep locked in people’s heads.
Where does this information actually need to sit?
This is the part most coverage skips over, so it is worth being clear. The knowledge that fuels these answers does not live in a separate ad setting somewhere. It lives on your product pages, your service pages and, right now especially, in your product data feed. That is where Google draws the raw material from.
At the moment the format leans heavily towards retail, because product data feeds are already structured, detailed and machine-readable in a way that most service pages are not. A lot of what is happening today is feed-driven: rich product data flowing into Performance Max campaigns, and now AI Max for Search, which then gives Gemini something specific to build an answer from. If your feed only carries a title, a price and a thin one-line description, the AI has very little to work with when a buyer asks something detailed.
My expectation is that this widens over time. Google is clearly moving towards using on-page content, your actual product and service descriptions, to fuel these conversational answers rather than relying on the feed alone. That is good news for service businesses, who cannot lean on a product feed in the same way, but it does put the onus back on the quality of the writing on the page.
The practical takeaway is straightforward. Your product and service descriptions need to contain the answers to the questions buyers are actually asking. Not marketing gloss, but the specifics: dimensions, compatibility, materials, timescales, what is included, what is not, and the common queries that come up before someone buys or enquires. If a buyer would ask it, the answer should already be written down somewhere Google can reach.
What does this mean for retailers versus service businesses?
If you sell products, your priority is the feed and the product pages behind it. Make sure each description answers the real questions buyers ask, and that your feed carries as much useful, accurate detail as the format allows rather than the bare minimum. If you sell services, you cannot rely on a product feed, so the weight falls on your service pages. Write them so they genuinely address buyer questions, because as Google leans more on on-page content, well-written service pages become the equivalent of a strong product feed.
How do you prepare your business to win with conversational ads?
The mechanics of the ad format are Google’s problem. The quality of what it has to work with is yours. So the practical work sits in getting your knowledge onto your pages and into your feed in a usable form rather than in fiddling with bid settings.
Start by auditing the questions your buyers genuinely ask. Not the keywords in your account, the actual questions, the ones that come up on sales calls and in support tickets. These are the queries a conversational format will need to answer, and you almost certainly have a record of them already if you go looking.
Next, make sure the answers to those questions actually appear where Google can use them: in your product descriptions, on your service pages, and in your product data feed. This is where the unique insight only your business holds becomes an asset an algorithm cannot scrape from anywhere else. If you sell precision engineering, the difference between a tolerance and a specification is not pedantry, it is the thing that tells a buyer you actually understand their world. Google’s earlier confirmation that Gemini powers the conversational experience in Google Ads, covered in the Adweek report on Google’s Gemini ads, makes clear this is now the direction of the platform, not a side experiment.
Then keep a human in the loop. AI-generated answers still need review, both for accuracy and for tone. A single wrong claim in a technical or regulated field is a liability, not a rounding error. The businesses that win here will be the ones that combine the speed of AI generation with the judgement of someone who knows the subject well enough to catch an error before it reaches a buyer.
A simple sequence to follow
- Gather the real questions buyers ask, from sales and support records. 2. Write clear, specific answers grounded in your genuine technical knowledge. 3. Put those answers where Google can reach them, in your product descriptions, service pages and product data feed. 4. Review every AI-generated output for accuracy and tone before it goes live. 5. Track which answers actually drive enquiries, and refine the weakest ones. None of this is glamorous, but it is the difference between an ad that answers a question and one that dodges it.
Frequently Asked Questions
What are Google’s Conversational Discovery ads?
They are a Gemini-powered ad format Google began testing in 2026 that answers a user’s specific question inside AI-driven search experiences. Instead of a static headline, the ad uses Gemini to build creative tailored to the individual search, drawing on the advertiser’s own expertise to form part of a genuinely useful answer.
Where should the information that fuels these ads sit?
On your product pages, service pages and, right now especially, in your product data feed. Google draws the raw material from there, so your descriptions need to answer the questions buyers actually ask. The format currently leans on feeds in Performance Max and AI Max, but is moving towards using on-page content too, which makes well-written pages increasingly important.
Do I still control what my ads say if AI generates them?
Yes. Google frames the conversational experience as a tool to combine your expertise with its AI, not to replace your judgement. It helps build tightly themed ad groups and optimised responsive search ads, but the underlying knowledge and the final review remain yours. Keeping a human in the loop is essential, particularly in technical or regulated fields.
How many B2B buyers research before contacting a vendor?
A 2025 Responsive report found 90% of B2B buyers conduct research before first contact, and nearly two-thirds now use generative AI tools as much as or more than traditional search. This means much of your evaluation happens before anyone speaks to you, so the substance you surface in AI-mediated moments strongly influences whether you get shortlisted.
What information should I prepare for conversational ads?
Document the questions buyers genuinely ask, then write clear answers into your product descriptions, service pages and product feed. Cover integration requirements, implementation realities, dimensions and specifications, where your product fits and where it does not. This structured, on-page detail is the raw material the AI needs to answer a real question well rather than falling back on generic claims.
Is this format only useful for large companies?
No. If anything it favours smaller specialists with genuine depth over larger firms relying on brand budget. Conversational ads reward specific, accurate answers rather than reach or recognition, so a small business that truly understands its niche can compete for a buyer’s attention against far larger competitors on the strength of its knowledge alone.
Final Thoughts
The shift towards conversational, AI-generated ads is not something to fear if your business has real substance behind it. For years, deep technical knowledge sat trapped in a 30-character headline, doing far less than it could. This format finally lets that expertise do the work it was always meant to do, answering the specific questions buyers actually ask at the moment they are deciding who to trust.
The catch is that the AI can only draw on what you give it. Thin input produces thin answers, and generic claims will look increasingly hollow next to competitors who genuinely understand their subject. The businesses that prepare properly, by structuring their knowledge and keeping a human hand on accuracy and tone, will be the ones these formats reward.
