AI Overviews and the Future of Paid Search

The old assumption in Google Ads was simple. If you bid well, wrote strong copy, and kept conversion tracking clean, you could still buy meaningful visibility on the results page.

That assumption has broken.

When Google AI Overviews triggered a 68% drop in paid search click-through rates, falling from 19.7% in June 2024 to 6.34% by September 2025, the issue stopped being a minor layout tweak and became a structural change to search behaviour, as reported by Search Engine Land’s coverage of Seer Interactive’s analysis. For UK SMEs, that matters because fewer clicks don’t just mean less traffic. They change how your budget behaves, how your reports look, and which campaign types still deserve your money.

If you manage ecommerce, lead gen, or both, AI Overviews and the Future of Paid Search is no longer a trend piece. It’s an operating issue. The practical question now is not whether AI is changing search. It’s how quickly you adapt your account structure, budgets, and measurement so you don’t keep paying more for declining visibility.

The Paid Search Landscape Is Changing Forever

A 68% drop in paid CTR on queries that trigger AI Overviews is not a creative problem or a bid problem. It is a change in how the search results page distributes attention, and UK SMEs need to treat it as a budget and measurement issue.

The practical risk is simple. You can keep the same spend, hold roughly the same average CPC, and still get less from search because fewer people reach your ad in the first place.

The old top-of-page model is under pressure

For years, many Google Ads accounts were built around a straightforward assumption. Strong ad rank plus clear intent produced a reliable flow of clicks. AI Overviews weaken that model because Google now answers part of the query before the user decides whether to click anything at all.

That matters most for businesses that rely on high-volume non-brand searches to fill the top of the funnel. In ecommerce, that often means category and comparison terms. In lead generation, it usually hits early research queries first, especially where users want definitions, options, prices, or quick summaries.

The result is not just lower traffic. It is harder budget allocation.

A marketing manager looking only at blended account performance can miss what is happening underneath. Brand may still hold up. Shopping may stay efficient for a while. High-intent generic terms may look acceptable on last-click reporting. Meanwhile, upper-funnel search campaigns start losing reach, remarketing pools shrink, and the cost of generating future demand rises.

Why this feels different from previous Google changes

Google has changed match types, bidding controls, shopping formats, and reporting many times. Those changes affected how advertisers competed in the auction. AI Overviews change what happens before the auction wins attention.

Users can now get enough information from the results page to delay the click, refine the search, or switch channel entirely. That creates a real trade-off for SMEs. If you cut top-of-funnel spend too fast, you protect short-term efficiency but reduce future demand. If you leave budgets untouched, you may keep paying for inventory that no longer produces the same volume or quality of visits.

That is why the better question is not, "Are AI Overviews affecting us?" The better question is, "Which campaigns are losing useful visibility, and what do we change first?"

For busy teams, three account-level effects show up early:

• Click volume becomes less reliable on informational and mixed-intent queries, even if impression levels stay healthy.
• Attribution gets messier because users may research on Google, return later through brand search, direct, email, or paid social, and make the original search campaign look weaker than it was.
• Budget efficiency splits by campaign type because branded, bottom-funnel, and feed-led activity often holds up better than broad generic search.

For UK SMEs, the response needs to be phased. Ecommerce teams should review where Shopping, PMax, and branded search are still defending revenue, then isolate generic search terms that are losing role and margin. Lead gen teams should separate high-intent quote or enquiry terms from research queries, tighten conversion definitions, and watch lead quality rather than form-fill volume alone.

If you want the wider strategic context, our guide to how AI assistants will choose ads explains why this shift will affect more than ad placement.

Understanding AI in Paid Search A Primer for Marketers

Not all AI in Google Ads does the same job. That’s where a lot of confusion starts.

Some AI predicts. Some AI generates. If you mix those up, you’ll misread what’s happening in your account.

Predictive AI has been in your account for years

Smart Bidding is the familiar example. Target CPA, Target ROAS, and similar bidding systems use machine learning to estimate the likelihood of a conversion and adjust bids accordingly. They don’t write answers. They model probability.

That kind of AI is useful when the auction is volatile because it can react faster than a person reviewing bids by hand. But it still depends on the quality of your inputs. Bad conversion tracking, weak audience signals, or poor feed data will still produce poor outcomes.

Think of predictive AI as an analyst. It studies patterns and makes fast bidding decisions.

Generative AI is different

AI Overviews are powered by generative AI, not just predictive systems. The easiest way to explain this to a non-technical stakeholder is to picture a very fast research assistant. It can read a huge amount of web content, identify common themes, and produce a summary that sounds coherent to a user.

That’s what a large language model does in practical terms. It processes language at scale and generates text that responds to a prompt or query in context.

For paid search, the implication is important. Google is no longer just matching a keyword to an ad and ranking results. It is also interpreting the query at a much deeper level and deciding whether the user might prefer an instant answer over a click.

Why your old keyword logic can fail

A traditional PPC structure often assumes this sequence:

1. User searches.
2. Google shows ads and listings.
3. User clicks.
4. Landing page does the selling.

AI Overviews interrupt that path. The searcher may read a summary, compare brands mentally, and never click on that first query at all.

That changes how you should think about:

• Keyword intent because informational and commercial language are blending more often.
• Ad copy because generic feature-led messaging becomes easier to ignore.
• Landing pages because users who do click may arrive later in the decision process.
• Measurement because the first search may influence the sale without delivering the click.

Your account now sits inside a search environment where Google doesn’t just rank options. It increasingly interprets the question first.

This guide to generative engine optimisation is a useful companion if you’re trying to align PPC with the wider AI search environment.

A simple way to explain it internally

Use this distinction when briefing your team or leadership:

If you keep those categories clear, your decisions get sharper. You stop asking whether “AI” works, and start asking which AI system is affecting which part of your funnel.

The Immediate Impact on Your PPC Campaigns

UK SME advertisers should expect more pressure on paid search efficiency as AI Overviews expand across more query types. The early signs are usually practical rather than dramatic. Click volume softens before revenue drops. Generic campaigns look less productive. Reporting starts to look inconsistent by search intent.

Lower visibility is usually the first warning sign

Semrush’s AI Overviews study projects that by November 2025, AI Overviews will appear on 15.69% of UK search queries. If that projection proves directionally right, fewer searches will send the same level of traffic through the standard ad-and-organic click path.

That matters because many SME accounts rely on steady volumes from non-brand discovery terms. Once Google answers more of those queries on the results page, your ads have to compete for attention after the user has already seen a summary.

In live accounts, this often shows up in ordinary ways:

• Stable spend with lower click volume
• Weaker support from generic campaigns into brand and remarketing
• Less consistent lead quality from early-stage searches
• More variance by device, query theme, or match type

None of that means the account is broken. It means the search results page is doing more of the journey before the click.

Tighter inventory usually means higher CPC pressure

Paid search gets harder when there are fewer high-value clicks available for the same commercial demand. That is the primary risk for SMEs with fixed budgets. You can hold spend flat and still lose ground if CPCs rise and exploratory traffic drops.

The pressure tends to hit these areas first:

• Head and mid-funnel terms, where AI-generated summaries are more likely to intercept attention
• Auction stability, because weekly manual bid changes struggle to keep up with faster shifts in query behaviour
• Commercial comparison searches, where users may get enough information from the SERP to delay or skip the click
• Shopping and broad targeting efficiency, especially if campaign structure and feed quality are only average

For ecommerce brands, category and comparison intent often weakens before high-intent product searches do. For lead generation, research-led terms usually lose value first, even when bottom-funnel service terms still convert well.

That trade-off matters for budget planning. Chasing lost traffic at any cost can protect volume in the short term while eroding margin.

Old KPI reading is less reliable on its own

CTR, CPC, conversion rate, cost per lead, and ROAS still matter. They just need more context than they did before.

A falling CTR does not automatically mean the ad is poor. It may mean the page layout changed and the click is harder to win. A campaign with strong ROAS may be harvesting lower-funnel demand while prospecting weakens. A stable CPL can hide a future pipeline problem if top-of-funnel traffic has fallen away.

I’d review account health through three practical lenses:

This is also the point where many teams discover their reporting is too shallow. If you only review blended account averages, you will miss where the damage starts. Segment by brand vs non-brand, new vs returning users, query intent, and campaign type.

Here’s a useful explainer if you want to see the broader mechanics in action:

What this looks like for ecommerce and lead gen

The commercial impact is different by model, so the response should be different too.

Ecommerce

Retailers usually feel the hit first on product discovery, category terms, and comparison searches. If the feed is messy, product titles are weak, or campaign structure lumps unlike products together, AI-shaped results make those issues more expensive.

The accounts holding up better usually do three things early:

• Improve feed data with clearer titles, better attributes, and tighter categorisation
• Separate products by margin and intent so budget does not drift into low-value inventory
• Review search term behaviour by product group instead of judging Shopping as one block

Common mistakes are predictable:

• Increasing budget to recover lost traffic without checking margin impact
• Using last-click ROAS as the only decision metric
• Leaving outdated Shopping or Performance Max structures in place

If Performance Max is carrying more of your visibility, tighter controls matter. This guide on how to control Google’s automated campaigns in Performance Max is useful if your ecommerce account has started to blur prospecting, branded demand, and low-quality reach.

Lead generation

Lead gen accounts tend to lose ground first on research-heavy searches. Those queries used to introduce the prospect. Now Google may answer enough of the question to reduce the click, or delay it until the buyer is further along.

The stronger response is usually operational, not cosmetic:

1. Tighten lead qualification signals inside the account.
2. Separate research terms from high-intent service terms.
3. Build remarketing paths that reflect a longer decision cycle.
4. Feed offline conversion data and CRM outcomes back into bidding where possible.

If your dashboard says traffic is down and CPL is flat, treat that as a warning, not reassurance. You may be preserving efficiency by shrinking future demand.

Current AI-Powered Tools in Your Google Ads Account

Google has already folded AI into the core tools many SMEs use every day in Google Ads. The practical question is not whether to use automation. It is where it improves coverage and efficiency, and where it needs tighter controls to protect margin and lead quality.

Performance Max is now part of the visibility layer

Performance Max is no longer a side test for spare budget. For many UK advertisers, it is now one of the routes into search demand that no longer behaves like a standard list of ten blue links.

That creates a real trade-off. Performance Max can widen reach and help you appear in newer placements, but it also reduces the amount of query and placement-level control many teams are used to. For ecommerce brands, that can mean stronger product exposure if the feed is clean. It can also mean wasted spend if low-margin lines, weak images, or vague titles are left unchecked. For lead generation, the risk is different. Volume can rise while sales quality slips.

Use Performance Max with clear structure:

• Group assets by product range or service line, not by broad business category
• Use audience signals based on real customer data, not generic interest segments
• Keep feed data accurate and specific for ecommerce, especially titles, pricing, availability, and GTINs
• Match landing pages closely to the offer so Google does not send paid traffic to weak or irrelevant pages
• Review search themes, brand split, and geographic performance regularly to catch drift early

If you need a tighter operating model, this guide on how to control Google’s automated campaigns in Performance Max is a useful reference.

Broad match works differently with strong signals

Broad match still gets dismissed because many advertisers remember how much budget it used to waste. In some accounts, that criticism is still fair. Broad match without clean conversion data is expensive guesswork.

Used properly, it serves a different role now. Search behaviour is getting messier, longer, and more conversational. Exact and phrase match still matter, especially for high-intent commercial terms, but they do not cover the full range of queries that can lead to a sale or qualified enquiry. Broad match helps Google find those variants, provided the account has enough guidance.

That guidance usually comes from four places:

• Primary conversions that reflect real business value
• Negative keyword lists that are actively maintained
• Revenue imports or value-based bidding rules where possible
• Campaign segmentation based on margin, service priority, or lead quality

For a small ecommerce business, that might mean separating high-margin ranges from low-margin catalogue traffic before testing broad match. For lead gen, it often means splitting qualified demo requests from softer actions such as brochure downloads or time-on-site goals.

Smart Bidding is only as good as the signal you feed it

Smart Bidding can improve efficiency. It can also scale bad decisions faster than manual bidding ever could.

I still see SME accounts using Target CPA or Maximise Conversions while counting every form fill, repeat lead, and spam enquiry as equal. In that setup, Google does exactly what it has been asked to do. It just does not help the business.

Before increasing reliance on automation, check the inputs:

For UK SMEs, the gap often becomes evident. The issue is rarely that Google’s AI tools are unavailable. The issue is that the account structure, tracking setup, CRM feedback loop, and reporting standard are not ready for them. Ecommerce businesses need product and profit data in better shape. Lead gen businesses need offline conversion imports, lead scoring, and a firmer definition of what counts as a valuable lead.

PPC Geeks supports that work through audits, feed optimisation, conversion tracking checks, and reporting across Google Ads, Microsoft Ads, Amazon, and paid social. That is useful if the problem is not one campaign type, but wasted spend across the whole acquisition mix.

Adapting Your Strategy for Near and Mid-Term Trends

The next stage of paid search won’t be won by bidding alone. It will be won by advertisers who connect campaign mechanics with brand authority and better measurement.

Ads inside AI Overviews change the job

Adthena data projects a 20 to 40% relative drop in paid search CTR due to AI summaries, and Google is mandating Performance Max and the new AI Max for Search campaigns to enable ad integration directly into AI Overviews, as covered by Search Engine Land’s reporting on AI Overviews accelerating change in paid search.

That creates an awkward but important trade-off.

On one side, advertisers want control. On the other, the routes into emerging placements increasingly run through automated campaign types. The practical answer isn’t to choose one extreme. It’s to decide where automation deserves freedom and where it needs constraints.

Brand authority starts affecting paid outcomes

As search becomes more answer-led, visibility doesn’t depend only on who bids. It also depends on whether Google can understand, trust, and surface your brand in context.

That’s where Generative Engine Optimisation, or GEO, becomes useful. In plain English, GEO means preparing your site, content, data, and brand signals so AI systems can interpret them cleanly and cite them with confidence.

For paid search teams, that matters because the user may first encounter your brand in an AI-generated summary, then click later through branded search, remarketing, Shopping, or another channel. If your business is absent from those summaries, the paid campaign inherits a weaker starting point.

What to build over the next 6 to 18 months

The strongest medium-term strategy is not “do more AI”. It’s more disciplined than that.

Focus on these capability areas:

• Measurement resilience
  Move beyond surface metrics. Bring CRM outcomes, qualified lead stages, and first-party signals into platform decisions where possible.

• Search and content alignment
  Paid teams and content teams need a shared view of priority topics, service language, and commercial proof points.

• Feed and landing page readiness
  Ecommerce brands should treat feeds as strategic assets, not admin files. Lead gen teams should build pages that answer questions clearly before asking for the form fill.

• Creative built for mixed intent
  Research-stage searchers need reassurance, not just an aggressive CTA. Bottom-funnel users still need speed and clarity.

Strong paid accounts in this environment don’t rely on a single click path. They create multiple ways for a buyer to recognise the brand, trust it, and return.

What won’t age well

Some habits are likely to underperform as AI-led search expands:

1. Judging campaign quality mainly by CTR
2. Separating PPC and SEO planning completely
3. Using broad automation without conversion discipline
4. Ignoring first-party data until attribution becomes a crisis
5. Treating branded search growth as proof the whole funnel is healthy

The near-term future belongs to accounts that can tolerate ambiguity without becoming sloppy. That means tighter tracking, clearer business signals, better site structure, and campaign types designed for AI-shaped SERPs rather than legacy ones.

A Phased Implementation Roadmap for UK SMEs

Most SMEs don’t need a dramatic rebuild in one week. They need a clear order of operations.

That matters even more because there’s still a gap in UK-specific SME benchmarking, while AI Overviews are appearing in 13 to 19% of searches internationally, which strengthens the case for first-party data integration and channel diversification such as Bing and Amazon, according to The USIM’s review of AI’s impact on paid search.

When external benchmarks are thin, your own data becomes more important. That’s why a phased plan works better than reactive tinkering.

The roadmap

Phase 1 foundations

Start with what your account is telling Google.

If conversion tracking is messy, every smart bidding decision becomes less trustworthy. If your feed is weak, Shopping and Performance Max lose relevance. If your landing pages are vague, expensive clicks still won’t convert.

For the first phase, prioritise:

• Conversion audit
  Make sure primary actions reflect real business value. Lead gen teams should separate high-quality enquiries from low-intent form fills where possible.

• Campaign inventory review
  Identify which campaigns rely most on informational and mid-funnel query types. Those are often the first to feel SERP disruption.

• Performance Max assessment
  Decide whether it’s structured well enough to keep, needs rebuilding, or needs tighter exclusions and asset grouping.

• Search term and feed review
  Ecommerce brands should look closely at titles, attributes, and product categorisation. Lead gen advertisers should review query language and landing page alignment.

A useful operational reference at this stage is this article on AI PPC with OpenAI ad services, especially if you’re trying to translate broad AI discussion into practical account decisions.

Phase 2 adaptation

Once the account basics are under control, move to structural adaptation.

This is the point where many teams either overreact or stall. The better move is selective change.

For ecommerce

Build campaign logic around commercial reality, not platform convenience.

• Push stronger products and categories into clearer campaign groupings.
• Test asset quality and feed refinement before changing every budget.
• Use remarketing and audience layering to recover users who researched but didn’t buy on first contact.

For lead generation

Improve qualification before chasing more volume.

• Feed CRM outcomes back into campaign optimisation where possible.
• Split early-stage query themes from service-led, high-intent terms.
• Build nurture paths for users who won’t convert on the first click.

Don’t use AI-led volatility as an excuse for a full account reset. Keep what still works. Replace what clearly doesn’t.

Phase 3 future-proofing

Here, your search strategy starts to look less like channel management and more like demand infrastructure.

The long-term priority is owning more of your signal set:

• First-party data from forms, calls, CRM stages, and customer lists
• Content and site structure that helps AI systems understand your expertise
• Cross-channel testing so Google isn’t the only place your demand capture relies on
• Shared reporting across paid search, ecommerce platform data, and sales outcomes

For SMEs, this doesn’t require enterprise complexity. It requires consistency. A modest account with clean signals, sensible automation, and disciplined diversification will usually outperform a larger account that keeps optimising for a SERP that no longer exists.

Your Next Steps in the AI Era

AI Overviews haven’t killed paid search. They’ve changed what competent paid search looks like.

The advertisers who struggle most will be the ones who keep reading the same dashboard in the same way. The ones who adapt will treat search as a wider visibility system, not just a click-buying machine.

That means a few practical shifts. Tighten conversion tracking. Use automation where it earns the right to be trusted. Improve feeds and landing pages. Bring paid, content, and first-party data closer together. Stop judging campaign health on CTR alone.

For UK SMEs, the opportunity is still real. A lot of competitors will respond too slowly, or they’ll hand too much control to Google without enough oversight. That creates room for businesses that stay disciplined.

If your account hasn’t been reviewed through the lens of AI Overviews and the Future of Paid Search, now is the time to do it. A serious audit should tell you which campaigns are exposed, which metrics still matter, and where your next pound is most likely to produce growth rather than drift.

Frequently Asked Questions

Will my ads appear directly inside AI Overviews

They can, but eligibility is increasingly tied to the campaign types Google is using for these placements. Right now, that means advertisers should pay close attention to Performance Max, Shopping, App campaigns, and AI-led matching approaches. If your account only relies on older, tightly manual structures, you may miss some of the emerging visibility opportunities.

Should I reduce my Google Ads budget because of AI Overviews

Not automatically.

Cutting budget without diagnosing the account usually makes things worse. The better move is to reallocate. Protect the campaigns that still capture strong intent. Review weaker query classes. Test diversification into channels such as Microsoft Ads or Amazon where relevant. For many SMEs, the problem isn’t total budget size. It’s budget distribution against changing search behaviour.

Which KPIs matter most now that CTR is less reliable

CTR still has value, but it can’t carry the full story.

Focus more heavily on:

• Qualified conversions
• Cost per qualified lead or sale
• Revenue or pipeline value
• Search term quality
• Branded lift and remarketing performance
• CRM-confirmed outcomes where available

If a campaign loses CTR but keeps producing profitable sales, the answer may be to refine rather than pause. If CTR looks acceptable but lead quality falls, the campaign may already be underperforming despite the headline numbers.

If you need a clear view of how exposed your account is to AI-led SERP changes, PPC Geeks can audit your tracking, campaign structure, feeds, and budget allocation, then show where to tighten performance and where to adapt for the next stage of paid search.