If you run Google Search ads, you are about to buy space inside a conversation.
Not a neat list of blue links. Not a tidy set of keyword buckets. A chatty AI Mode flow where the user asks, compares, refines, and keeps going.
Google is also pushing new Gemini-built ad formats into Search, including AI Mode surfaces. Here is the official update: New ad formats built with Gemini coming to Google Search.
The annoying part
Creative and bidding still matter, sure. But if your offer pages are fuzzy and your proof is thin, AI Mode will not save you. It will just send you more expensive confusion.
Also, reporting is not going to get magically cleaner. Google Ads says ads can appear in AI Overviews and those clicks show up in normal reporting. No special segment. Source.
So you need your own lead quality signal. Fast.
The quick playbook
Do this before you touch your bid strategy.
- Build one real offer page per service. Who it is for, what you do, what it costs (even a range), and what happens next.
- Turn proof into assets. Numbers, screenshots, photos, before/after, outcomes, names. This helps: Case Studies That Close Deals.
- Get UTMs under control. If you cannot connect click to lead to revenue, you are buying vibes. UTMs Without the Tears.
- Grade leads. Every lead gets tagged: Qualified, Not Qualified, Spam. Then you optimize on that.
- Fix response time. If you are paying for intent and answering tomorrow, stop. The Lead Response System.
What Google actually announced
Google Marketing Live 2026 was a lot of AI talk. The advertiser-relevant bits are the formats and placements.
Google says it is testing or rolling out formats like:
- Conversational Discovery ads in AI Mode.
- Highlighted Answers, which can surface a sponsored answer inside the experience.
- Direct Offers, which can pull an offer into the flow.
Google’s own write-up: New ad formats built with Gemini coming to Google Search. Also relevant context: Google started testing Conversational Discovery ads in AI Mode earlier this year. Source.
Start with your inputs
AI Mode is a decision assistant. It compresses the research phase. That means your landing page has less time to earn trust, and more work to do when it gets the click.
Make your business easy to recommend, easy to verify, and easy to pick.
Input 1: Offer clarity
If your page reads like a brochure, expect curiosity clicks. Then a quiet bounce.
Build one strong offer page per service line. Make it blunt:
- Who it is for (and who it is not for).
- What you actually do.
- Typical timeline.
- Price range or starting point.
- Two to three proof points with specifics.
- A single primary next step. Call, form, booking link. Pick one.
Input 2: Proof that survives a skeptical buyer
Proof beats adjectives. Give the buyer something they can repeat in an internal meeting without feeling dumb.
If you need a structure, steal this: Case Studies That Close Deals: A Simple Proof Asset Template.
Input 3: Measurement that tells the truth
If you optimize to one default conversion, you will eventually learn how to buy junk leads at scale.
Two notes worth remembering:
- Google Ads says AI Overview ad clicks appear in your standard reporting, and reporting is not segmented specifically for AI Overviews. Source.
- Early research on AI Overviews suggests click patterns are shifting and attribution gets messy. If you want the academic version, see: The Great Decoupling: Search to Shopping in the Era of AI Overviews.
Translation. You need your own lead quality signal.
Lead quality: the missing conversion
Good lead, bad lead. That is the game.
Simple system:
- Every new lead gets tagged within 48 hours: Qualified, Not qualified, Spam.
- Qualified gets one sentence of why: budget, fit, urgency, geography, whatever matters to you.
- Once a week, you pull a list and look for patterns.
This pairs with fast follow-up. If response speed and routing is a mess, start here: Stop Losing Leads After They Click.
What to ignore
- “AI Mode keyword lists.” You need offer clarity and proof, not 600 new keywords.
- City-swapped landing pages. They will not save you. They might hurt you.
- “Just turn on everything.” If you cannot measure lead quality, you are buying randomness faster.
Where SigServe fits
Paid Search is a system: offer, proof, measurement, and follow-up. When any of those is weak, you pay to learn that it is weak.
If you want help tightening the inputs that make these AI-driven placements work, this lives under Digital Strategy & Advertising. For the bigger shift, start here: Google Search Is Becoming an Action Engine.
Want cleaner leads from Search?
We can help you tighten the offer, build proof assets, and set up measurement that tells the truth.