Key takeaways
- Google is testing exact and phrase match Search campaign text ads inside AI Mode when intent is explicit and direct.
- The main advertiser risk is blended reporting, where AI Mode traffic changes spend and CPA without a clear decision trail.
- Restrictive match types are not enough protection if conversion tracking, negatives and campaign budgets are weak.
- UK advertisers should segment keywords by intent, tag AI-style search terms and import qualified conversion data.
- AI Max and Performance Max still matter for broader conversational searches, but standard Search now needs tighter controls too.
AI Mode Search ads put a new pressure point into accounts that already struggle with query control. Google is testing standard Search campaigns inside AI Mode, and that matters because exact and phrase match advertisers will no longer be dealing only with the conventional search results page. The same keyword structure that once bought you controlled intent now has a route into an AI-led experience.
For UK advertisers, the risk is not that every campaign suddenly changes overnight. The risk is quieter and more expensive: your existing Search campaigns start picking up traffic from a new surface, reporting lags behind the way budgets move, and Smart Bidding reacts before your team has separated good intent from noise. That is why our recent guidance on Google Ads automation signals matters here. Automation works only when the inputs and controls are clean.
Our view is simple. Treat this as a paid search control issue, not a novelty AI feature. If AI Mode becomes another route for standard Search ads, advertisers with disciplined keyword builds, clean conversion data and tight budgets will gain reach. Sloppy accounts will buy extra ambiguity.
What changed with AI Mode Search ads
Google has started a small experiment that allows traditional Search campaigns using exact and phrase match keywords to serve text ads in AI Mode. That is the core change. Advertisers do not need AI Max or Performance Max to appear in every eligible AI Mode scenario during this test.
The eligibility is narrow. Google has said exact and phrase match keywords are eligible only where its systems identify explicit and direct user intent. That means the query must map clearly to the advertiser’s keyword. This is not broad conversational matching under a different label.
There is also a format distinction. Standard Search campaigns are being tested for text ads. Google is still positioning AI Max and Performance Max for Search as the route into more complex AI Mode opportunities and newer formats such as Highlighted Answers.
Why AI Mode changes the money flow
The commercial mechanism is straightforward. If an existing exact match keyword starts serving in AI Mode, budget moves without the advertiser rebuilding the campaign. The campaign remains familiar, but the context of the click changes. A user has interacted with an AI search experience rather than a standard results page. Their level of comparison, certainty and urgency will differ from a conventional keyword click.
That matters because bidding systems do not care where your board expected the money to go. They learn from conversion signals. If AI Mode clicks convert well, Smart Bidding will lean in. If they convert poorly but still look close to converting, the system will keep paying while it gathers data. The account owner sees spend, clicks and CPA movement. They do not get a clean label that says this cost increase came from a new AI Mode serving path.
The accounts most exposed are not the ones running messy broad match alone. They are the accounts that believe exact and phrase match are a complete safety net. Exact match already moved away from old-school literal matching. Phrase match now works across meaning, not just word order. Add AI Mode interpretation on top and the practical question becomes: did Google understand the commercial intent the way your sales team would?
Query matching becomes a finance problem
When a phrase keyword picks up a query in AI Mode, the ad still has to justify the cost against the user’s intent. A high-intent phrase such as emergency boiler repair has a different value from a longer AI-style prompt asking whether a boiler repair is worth it before selling a house. Both contain useful words. Only one is likely to produce a profitable lead quickly. This is where a solid competitive bidding framework earns its keep.
This is why wasted keyword spend matters more now. In our UK Google Ads spend report, PPC Geeks keyword-CPA analysis across 49 UK accounts in Q2 2026 found the median account put about 24% of search spend into keywords that were not pulling their weight on last-touch, meaning they converted nothing or cost more than three times that account’s own average. That figure rests on each account’s own conversion data, which is distorted in many accounts, and some of the spend is legitimate brand defence, testing or assist activity. The like-for-like direction still matters: in the 42 measurable accounts across both quarters, the median rose about 2 points, with 20 accounts getting worse and 13 improving. Accounts with trustworthy tracking showed a median of 36%, so the headline number is not alarmist.
AI Mode Search ads will amplify that weakness. A keyword that already attracts borderline traffic in standard Search gets another surface to spend through. If your conversion tracking is loose, the system cannot tell the difference between a lead worth pursuing and a form fill that sales will reject within five minutes.
PPC Geeks’ View
The specific problem advertisers will face is AI Mode traffic being blended into normal Search decision-making. If Google does not give advertisers clear segmentation quickly, teams will judge keywords using mixed context data. That leads to poor decisions. You pause a keyword that still works on the standard results page, or you increase budget because blended CPA looks acceptable while AI Mode is dragging down lead quality.
We see the biggest risk in lead generation accounts using exact and phrase match with Smart Bidding, especially where offline conversion imports are late, incomplete or absent. The campaign looks controlled because the match types are restrictive. The sales outcome tells a different story: more enquiries, weaker qualification, and less clarity on which queries deserve the next pound.
Restrictive match types are not the same as restrictive economics. If Google adds a new serving surface, advertisers need proof that the extra clicks are profitable, not just eligible.
— Mark Lee, Senior Account Manager, PPC Geeks
The immediate takeaway is clear. Do not wait for a full rollout to tidy the account. Build the reporting, negatives and conversion quality checks now, while the test is still narrow. This is exactly the type of issue we look for in a free Google Ads audit, especially where automation, tracking or campaign structure is affecting performance.
What advertisers should do next
Do these checks before AI Mode Search ads become a routine line in your weekly budget conversation.
- Segment exact and phrase match spend by intent tier. Build a simple keyword map with three groups: urgent commercial intent, research intent and mixed intent. Put budget caps or tighter CPA targets around the mixed group. Do not let the same bid strategy treat all phrase match traffic as equal.
- Export search terms every week and tag AI-style queries manually. Look for longer, question-led and comparison-led terms that would fit an AI Mode journey. Add labels in a spreadsheet: buy now, qualify first, research only, poor fit. Feed the poor-fit group into negative keyword work.
- Separate campaign budgets where the economics differ. If one campaign contains emergency leads, planned projects and advice-led terms, split them. AI Mode will make mixed campaigns harder to read because the same budget is buying different states of intent.
- Fix conversion quality before raising targets. Import qualified lead, booked appointment, closed sale or revenue data into Google Ads. A form submission alone is too weak. If Smart Bidding receives soft conversions, it will optimise towards soft outcomes. Our guide to why conversion tracking matters for PPC covers the setup.
- Build a dashboard view for serving-surface anomalies. Create a report that compares keyword CPA, conversion rate, search term length, impression share and lead quality by week. Our guide to PPC dashboard creation gives the structure: show the decision, not every metric.
- Run an AI Max test separately from standard Search. If you want to compare Google’s automated AI route with controlled keyword campaigns, run a contained experiment rather than merging everything. Use one product line or lead type, fixed budget, shared conversion goals and a 30-day readout. Our AI Max experiments guide explains how to stop the test from contaminating the rest of the account.
There is one reporting discipline we would add for every UK advertiser: annotate your account now. Add a note for the week you first see unusual exact or phrase match movement, especially if query length, CTR or lead quality changes. Annotation will not solve attribution, but it gives your future analysis a fixed point instead of a guess. If you would rather hand this to a team that runs these checks daily, that is what our Google Ads agency does.
What this means for your campaigns
AI Mode Search ads are not a reason to abandon standard Search. They are a reason to make standard Search sharper. The advertisers who benefit will be the ones who know which exact and phrase keywords deserve extra reach, which ones need negatives, and which ones exist only because legacy structure was never cleaned up.
The original Search Engine Land report on traditional Search in AI Mode confirms this is a small test tied to exact and phrase match where Google sees explicit user intent. Google’s own Marketing Live 2025 announcement shows the wider direction: ads are being woven into AI-led search experiences, not kept in a separate PPC box. Google’s documentation on keyword match types is worth rereading too, since exact and phrase behaviour is the hinge of this whole test.
That is the key point. Your budget controls, conversion imports and search term process need to match the way search is changing. If they do not, the account will still spend. It will just give you less certainty about why.
If you would like a second pair of eyes on how this affects your account, our team offers a free Google Ads audit, with no strings.
Frequently asked questions
What are AI Mode Search ads?
AI Mode Search ads are standard Search campaign text ads being tested inside Google’s AI Mode experience. In this test, exact and phrase match keywords can serve ads when Google identifies explicit and direct user intent.
Do advertisers need AI Max or Performance Max for AI Mode?
Not for every tested placement. Google’s current experiment gives traditional exact and phrase match Search campaigns a limited route into AI Mode. AI Max and Performance Max remain the products Google is positioning for broader conversational searches and newer AI ad formats.
What is the biggest risk for UK advertisers?
The biggest risk is budget moving into a new AI-led surface without clean reporting separation. That makes keyword CPA, lead quality and Smart Bidding decisions harder to interpret.
Should advertisers change match types now?
Do not rush into broader match types because of this test. First, audit exact and phrase match performance, tighten negatives, split mixed-intent campaigns and fix conversion quality.
How should lead generation accounts prepare?
Lead generation accounts should import qualified lead stages, not just form fills. Smart Bidding needs sales-quality signals, otherwise it will optimise towards enquiries that look cheap but fail commercially.
