
Google Ads in AI Search: What Changes for SMBs
AI answers are changing search. What that means for Google Ads in smaller companies, which campaign types matter and what you can prepare today.
TL;DR
AI answers mainly shorten the path for knowledge questions; commercial searches, in my experience, still run through ad surfaces, and Google is additionally placing ads inside those AI surfaces. I am not aware of a separately bookable AI placement or a dedicated reporting segment for it: the surfaces are served from existing Search, Shopping, Performance Max and Demand Gen campaigns. For small and medium-sized businesses the order therefore stays the same, first clean conversion data, then complete assets and feeds, then reach.
Ever since AI answers started appearing above the search results, the same question comes up in almost every first call: are Google Ads still worth it if Google delivers the answer itself? My short answer from practice: yes, but the work shifts.
First, some framing so it is clear what this text is and what it is not. This is a practitioner's view from live accounts of small and medium-sized businesses, not a market forecast. I deliberately quote no percentages for click losses and no rollout dates I have not verified myself. Where I am estimating, I say so. Which AI tools I actually use in my own work is documented openly on the AI in numbers page.
What is shifting in search right now
Two things are visible, and you can verify both yourself by typing in your own core terms.
First, Google answers more and more questions directly at the top of the results page, as a summarised AI answer with source links. Second, there is a conversational mode alongside it, where users ask follow-up questions, compare and narrow things down instead of starting a new search.
The practical effect in the accounts I manage: the path to the answer gets shorter, the path to the website gets longer. Anyone asking a pure knowledge question clicks through less often. Anyone who wants to buy, hire or call still lands very quickly in an environment with ads. That environment is Google's business model, and I see no reason to assume Google will abolish it.
How strongly this hits you depends, in my experience, less on your industry than on the query type. Advice-heavy topics with many knowledge questions feel more of it, clear-cut needs such as a broken boiler or a short-notice appointment feel little. I do not have solid, cross-industry numbers on this, and anyone quoting you some should be able to back them up.
Ads inside and next to AI answers
Google shows advertising inside and around AI answers, labelled as sponsored. For day-to-day work, two points about that matter more than the format itself.
First: I am not aware of a placement target that can be booked for AI answers alone. So you do not create an AI campaign. These surfaces are served from what is already running, from Search, Shopping, Performance Max and Demand Gen campaigns. The lever is therefore not in booking but in supplying: how well does the system understand what you offer, for whom, at what price and in what location?
Second: I do not know of a reporting segment that cleanly separates which clicks came from an AI answer. If someone sells you an AI visibility report for Google Ads, ask which field in the account that number comes from. In practice this means you keep steering by conversions, revenue and search terms, not by a new metric.
Together these two points lead to an uncomfortable but honest consequence. Control keeps moving towards data and material. Your influence sits in what you hand the system.
What this means for the campaign types
Search campaigns. Query language is getting longer and closer to spoken sentences. Narrow keyword lists cover that increasingly poorly, broad match combined with Smart Bidding increasingly well. The price is spread. That is why the search terms report has become the most important weekly routine in my accounts, together with well-maintained negative keywords. The basics are in the keyword research guide.
Shopping and Performance Max. Here the product feed is the raw material. Titles, attributes, availability, price: whatever is missing there, no algorithm can guess. In my view feed quality is now part of ad quality, not a side task next to it.
Demand Gen. Demand often forms before someone knows your name or starts a search at all. Demand Gen addresses exactly that phase. For very small budgets it is not mandatory, but for offers that need explaining it is a sensible addition, as described in the guide to Demand Gen campaigns.
Brand campaigns. The underrated movement: someone reads an AI answer that mentions your company and then searches for your name. If you are not visible on your own name, you give away precisely the cheap click the AI answer just handed you. So brand campaigns are the last thing I would switch off, not the first.
Three pieces of homework that pay off regardless of format
The pleasant part of this transition: everything that prepares you for AI surfaces already improves performance today. There is no bet here that you can lose.
1. Get conversion data in order
The more automated delivery becomes, the more everything depends on the one signal you supply yourself: what counts as a conversion for you, and what is it worth? Concretely that means counting without duplicates, values instead of pure counts, working consent handling and, where possible, the return channel from your CRM for the enquiries that actually turn into jobs. If that part wobbles, the system reliably optimises for the wrong thing. How I set it up is described in the conversion tracking guide.
2. Feed your assets, feed and landing page
Ads are increasingly assembled from building blocks rather than delivered as a finished unit. For you that means enough headlines and descriptions that make sense on their own, images in several formats, a complete feed and a landing page that speaks the same language as the ad. A sentence that only works together with the one next to it is a risk in this environment.
3. Appear in sources worth citing
AI answers are assembled from sources. Anyone who does not appear in those sources does not appear in the answer either. So honest framing includes this: your own content that answers a question clearly and factually, a well-kept business profile, correct contact and location data and mentions where your industry happens are worth more today than they used to be.
So that no false expectation arises: I do not offer SEO. I mention this point because it affects the ads side, not because I am selling it to you. If you need support with it, a specialised partner is the right way to go.
What I am not claiming
- I am not claiming that click prices rise by a specific amount because of AI answers. There are arguments in both directions: less organic space suggests more pressure on paid surfaces, new surfaces suggest more inventory.
- I am not claiming that a specific share of searches is affected. I do not know that number.
- I am not claiming that I can place you inside AI answers. Nobody can commit to that.
- I am not claiming that classic search ads will disappear any time soon. It does not look that way to me.
What I am claiming: accounts with clean data and good material get through any format shift better than accounts without.
My assessment for small and medium-sized businesses
If you manage a budget in the range of a few thousand euros a month, AI search changes your priority list less than the headlines suggest. The order stays the same in my view: first be able to measure, then supply material, then buy reach.
What I would change is the routine. Once a month, look at what happens when you type in your three most important search terms yourself. Is there an AI answer? Does your company appear in it? Who gets named instead? That takes ten minutes and replaces every forecast with an observation.
And if you do not want to do this on the side: that is exactly what ongoing Google Ads management is for, from 1,000 euros a month. As an active Google Partner I hear about changes to formats and campaign types early, which helps while search is being rebuilt. A more detailed take on AI Overviews and AI Mode is in the article on AI search in 2026, and how much AI tools actually feature in my own work is documented on the AI in numbers page.
As of August 2026
Frequently asked questions
Do AI answers replace Google Ads?
In my practice it does not look that way. AI answers mostly handle knowledge questions and shorten the path there. Anyone who wants to buy, hire or call still ends up in an environment with ads, and Google is additionally building advertising into the AI surfaces. What changes is not whether you advertise but how: more automation, more weight on conversion data, assets and feeds. That is my assessment based on accounts I run, not a guarantee.
Can I book ads inside AI answers specifically?
I am not aware of a placement target that can be booked for AI answers alone. Google serves these surfaces from existing campaigns, mainly Search, Shopping, Performance Max and Demand Gen. I also do not know of a reporting segment that cleanly separates clicks coming from an AI answer. If someone promises you a targeted booking or an exact report for it, they should be able to show you where in the account that number comes from.
What should a small business tackle first?
In this order: first check conversion tracking, including values and consent handling, because any automation is only as good as that signal. Second, complete your ad assets, product feed and landing page so the system actually has material to work with. Third, search your own core terms once a month and note whether and how your company appears in the AI answer. All three pay off even if little changes about the AI surfaces.
Mijo Jurisic
Google Ads consultant & founder of MJ Marketing. Five-plus years of hands-on practice: from a self-taught start to the Google Premier Partner programme with 500+ direct Google Ads clients and β¬20M+ in managed media spend.
Share this article:



