Paid Search Is Going Top-Of-Funnel In AI Mode: How SMEs Should Rebuild Measurement, Landing Pages, And Budgets Before It’s Too Late
Google’s AI Mode is pulling search behavior earlier in the buying journey—yet Google admits AI Mode monetizes lower. That mismatch forces advertisers to prove discovery value without getting cheaper clicks. Here’s a practical playbook to restructure accounts, measurement, and content so upper-funnel queries become profitable (and defensible) in 2027.
Paid search is changing shape again—only this time it’s not just a new ad format or another automation toggle. It’s a behavior shift.
Google is telling us two things at once: people are using AI-powered search experiences to ask longer, earlier, more exploratory questions—and those AI Mode queries monetize at a lower rate than traditional search. Put together, that creates a planning problem for every advertiser heading into 2027: your inventory is moving up the funnel while your reporting, budgets, and landing pages are still anchored to bottom-funnel outcomes.
This editorial is written from my perspective as Marius Dosinescu at AYSA.ai. My bias is practical: if a platform shifts the playing field and won’t (or can’t) fully explain performance, businesses need systems that measure reality, adapt content and conversion paths, and execute changes quickly—with accountability and approvals.
Primary source and inspiration: Search Engine Journal’s coverage of Google’s comments about AI Mode and monetization (external link).
Concise summary
- AI Mode is pulling queries earlier in the buying journey. Those “research-first” searches create new paid opportunities, but they can depress last-click KPIs if you evaluate them like brand or high-intent search.
- Google has acknowledged AI Mode monetizes lower. That implies more query growth without a proportional increase in paid monetization—meaning advertisers may not receive a “discount” that offsets top-funnel inefficiency.
- Your defense is measurement + page experience. You need reporting that surfaces assists, segments long queries, and ties early Clicks to downstream outcomes—plus landing pages that answer follow-up questions so discovery becomes demand.
- Execution speed now matters as much as strategy. As AI-driven results evolve, businesses that can monitor changes and implement site updates quickly will compound advantages.
Key takeaways (what to remember when you close this tab)
- Stop treating “AI discovery traffic” like “purchase intent traffic.” Measure it differently, budget it separately, and route it to pages built for research.
- Segment queries by intent signals you can actually observe. Word-count is a crude but useful starting point; build additional segments using themes (comparison, “best”, “vs”, “for [use case]”, “how to choose”).
- Build landing pages for multi-turn behavior. AI experiences encourage follow-up questions; your pages must pre-answer the next steps (comparison, constraints, pricing, returns, proof, and fit).
- Assume the platform won’t give you perfect transparency. Create your own durable proof: holdouts, experiments, path reporting, and content-level engagement metrics.
- Unify PPC + SEO + content + CRO around one funnel story. In an AI-shaped SERP, the boundary between “paid click” and “organic influence” gets fuzzier—but the customer journey gets more explicit.
Table of contents
- The real shift: Search is expanding upward, while PPC is still judged at the bottom
- Why “AI Mode pays less” matters (even if you don’t advertise in it yet)
- What changed in search behavior: longer queries, follow-ups, and multimodal inputs
- The new ad dynamics: ads matched to the query and the AI response
- A measurement reset: Stop forcing AI discovery into last-click KPIs
- How to protect budgets: restructure accounts to isolate discovery risk
- Practical reporting: build a “long-query” segment you can defend
- Landing pages are the new keyword list: Win the follow-up, not just the click
- If you sell products: feed quality and Merchant Center fields become conversion levers
- A concrete SME scenario: “The clinic that bought curiosity”
- What agencies should rethink for 2027: deliverables, dashboards, and accountability
- Where AYSA.ai fits: Approved execution for a messy, cross-channel reality
- What to do next (action list)
- Sources and further reading
The real shift: Search is expanding upward, while PPC is still judged at the bottom
For years, the default mental model for search ads was simple:
- User has a need
- User expresses need as a query
- Advertiser bids on query
- User clicks
- User converts
That model was never universally true (research has always existed), but it was operationally useful because it aligned with how most PPC teams were evaluated: conversions, ROAS, CPA, and “what happened after the click.”
AI Mode and AI Overviews represent a new version of search where users are encouraged to:
- Ask longer questions rather than compress intent into keywords
- Refine and follow up in multiple turns
- Mix inputs (voice, images, documents) instead of only text
- Compare options earlier and more thoroughly
That naturally shifts more queries toward discovery and consideration—top and mid funnel. The “problem” isn’t that those queries exist. The problem is that most advertisers are still running budgets and reporting as if all clicks should behave like bottom-funnel demand capture.
When you force top-funnel behavior into bottom-funnel KPIs, you get predictable outcomes:
- Discovery clicks “fail” last-click conversion goals
- Smart bidding learns toward short-term conversion signals
- You unintentionally train the account away from emerging query patterns
- Finance sees declining efficiency and demands cuts
In other words, you create a self-fulfilling prophecy: “AI traffic doesn’t work,” when what actually failed is the measurement model and the post-click experience.
Why “AI Mode pays less” matters (even if you don’t advertise in it yet)
The most important line for advertisers isn’t about query length—it’s about monetization.
In the SEJ reporting, Google’s leadership indicated that AI Mode queries have lower monetization than traditional searches, at least right now (SEJ source).
You don’t need to be eligible for AI Mode ad placements to feel the effect. Why?
- If user behavior shifts toward earlier queries, the total mix of searches changes—even outside AI Mode.
- When platforms push experiences that monetize less, they tend to compensate elsewhere (more inventory, more formats, more automation, more “match types,” more pressure to broaden targeting).
- Advertisers often pay the same auction rates while intent quality becomes more variable.
So the “AI Mode pays less” idea isn’t just an investor story. It’s a warning label for practitioners: expect more exploratory clicks priced like commercial clicks. That’s the risk.
The opportunity is equally real: if AI Mode unlocks incremental queries—new questions people didn’t bother searching before—then someone will be the brand that earns trust during exploration and then captures the purchase later.
The question is: will that be you, or will you finance your competitors’ growth by spending on discovery without a system to convert it?
What changed in search behavior: longer queries, follow-ups, and multimodal inputs
In the SEJ piece, Google’s representatives described users entering longer, more detailed questions, with query behavior moving “up the funnel.” They also discussed follow-up queries and multimodal usage (voice, images, documents) within AI experiences (SEJ coverage).
Even without relying on specific numbers (because Google’s public reporting varies by context and timeframe), the directional trends are consistent with what many teams have already observed:
- Queries look more like prompts. Users include constraints (“for a small apartment,” “under $500,” “for sensitive skin,” “for winter travel”).
- Refinement is part of the experience. Users aren’t “done” after one search; they’re building a thread.
- Search is blended with evaluation media. People want reviews, videos, podcasts, comparisons, and situational proof.
This matters because PPC accounts and landing pages were optimized for the opposite behavior: short, transactional queries landing on category pages designed to push a decision quickly.
In AI-influenced search, the winning move is often to slow down and help the customer choose—because the customer will choose anyway. The only question is whether they’ll choose with you.
The new ad dynamics: ads matched to the query and the AI response
One subtle but consequential concept in Google’s public framing (as reported by SEJ) is that ads in AI experiences can be matched not only to the query, but to the content of the AI response itself.
That creates a new kind of relevance model:
- It’s not enough to target keywords; you need to be relevant to the answer space.
- Page quality and product feed depth can influence what the system can confidently connect.
- Brands that provide structured, specific, situational content are more “matchable” in these environments.
Practically, this increases the value of assets that historically lived in “SEO land” rather than “PPC land”:
- Clear product comparisons
- FAQ content grounded in real customer questions
- Trust and proof elements (policies, certifications, clinician bios, case studies, reviews)
- Structured data and clean Internal linking
In other words, AI ad placements don’t just reward bidding strategy. They reward businesses that have done the work to be understandable.
A measurement reset: Stop forcing AI discovery into last-click KPIs
If you do only one thing differently in the next quarter, do this: separate “discovery” from “demand capture” in how you measure and communicate performance.
Last-click ROAS is a blunt tool. It’s fine when user behavior is short and transactional. It’s misleading when behavior becomes threaded, comparative, and multi-session.
What changes in AI-shaped journeys
- The first click is less likely to convert.
- The first click is more likely to create a return visit via direct, email, brand search, or organic.
- The “value” of the first click is often learning, trust, and shortlist inclusion.
What you should measure instead (or in addition)
Without inventing new metrics, you can build a more defensible view using tools most advertisers already have:
- Conversion paths and assisted conversions (in whatever analytics stack you use). If discovery clicks show up earlier in paths, that’s evidence they’re not waste.
- New user rate and engaged sessions on discovery landing pages. Early clicks should show “learning” behavior: scroll depth, time on page, navigation to comparisons/policies, use of site search.
- Brand search lift (trend lines, not perfect Attribution). If discovery expands reach, brand demand should rise over time.
- Incrementality tests where possible: GEO tests, time-boxed holdouts, or campaign experiments.
Important constraint: attribution is messy, and AI experiences add more complexity. Don’t promise precision. Promise process. Your goal is to create durable proof that stands up in a budget review.
How to protect budgets: restructure accounts to isolate discovery risk
When platforms introduce new inventory with uncertain conversion behavior, blending it into your core campaigns is the fastest way to lose stakeholder trust. You need budget isolation and clear hypotheses.
Here’s a pragmatic way to structure this.
1) Create a dedicated “AI discovery” budget line
Whether you use broad match, keywordless targeting, or other automation-heavy approaches, isolate it with its own:
- Campaign(s)
- Budget
- Reporting view
- Landing page set
Why: if you run discovery inside your main non-brand conversion campaigns, you’ll either (a) kill discovery too early because it hurts CPA, or (b) let it quietly degrade the whole account and then get blamed.
2) Define the job-to-be-done and the acceptable payback window
Top-funnel spend needs a different contract with the business. Examples:
- “Acquire qualified first-time visitors to our comparison hub, at $X per engaged session.”
- “Increase qualified consult requests over 60–90 days by building consideration earlier.”
- “Drive product-page depth interactions for new categories where we lack brand awareness.”
If finance expects a 7-day ROAS on awareness queries, the test is dead on arrival.
3) Tighten guardrails, not just targeting
Common mistake: teams respond to top-funnel volatility by tightening keywords until they’re basically buying the same bottom-funnel terms as before.
Better: maintain reach but improve quality through guardrails like:
- Landing page eligibility (send exploration traffic only to pages designed for it)
- Negative keyword frameworks based on irrelevant use cases
- Geography, schedule, device adjustments based on business reality
- Creative that disqualifies bad-fit clicks (“For enterprise teams only,” “Serving Austin only,” “Not for emergency care”)
In AI-shaped search, disqualification is a form of optimization.
Practical reporting: build a “long-query” segment you can defend
One of the simplest, most defensible ways to detect the “moving up-funnel” effect is to segment queries by length. SEJ’s source suggested operationalizing this idea by exporting search terms and grouping by word count (reference).
Word count is not intent. But it correlates with specificity and research behavior often enough to be useful—especially for communicating with non-marketers.
How to build the segment (simple version)
- Export your search terms for the last 30–90 days.
- Add a column for word count.
- Create buckets like:
- 1–3 words (often navigational or transactional)
- 4–7 words (mixed intent)
- 8+ words (often research, constraints, “how to choose”)
- Compare each bucket on:
- CPA / ROAS (last click)
- Engagement on site
- Assists in conversion paths
What you’re looking for
- If long queries have low last-click performance but appear early in paths, they’re likely discovery that needs better nurturing (not immediate elimination).
- If long queries have low last-click performance and don’t appear in paths, you might be buying curiosity that isn’t relevant—time to tighten.
This segment becomes a conversation tool. It lets you say, “We’re buying a different kind of search now, so we’re measuring it differently.” That’s how you protect budgets while you learn.
Landing pages are the new keyword list: Win the follow-up, not just the click
In AI-influenced search, your landing page doesn’t just need to be relevant. It needs to be the natural next step in a multi-turn journey.
Here’s the hard truth: if someone searches a 20-word query with constraints and context, and you drop them onto a generic category page, you’re not “losing conversions.” You’re telling the user they were wrong to click.
What “top-funnel landing page” actually means
It’s not fluff. It’s specificity.
- “Which option is right for me?” guidance
- Comparison tables (not marketing adjectives)
- Constraints and fit (sizes, compatibility, contraindications, shipping limits, lead times)
- Proof (reviews, policies, guarantees, certifications)
- Clear next step: quiz, selector, consultation, sample, downloadable checklist
Build pages around “follow-up questions”
If AI encourages follow-up queries, treat your landing page as an answer to the next five questions the customer will ask. For example:
- “Will this work for my situation?”
- “What are the trade-offs?”
- “How much does it cost and what’s included?”
- “What do real customers like me say?”
- “What happens if it doesn’t fit / I need to return / I need support?”
In many SMEs, these questions are answered somewhere on the site—but scattered across PDFs, policy pages, or outdated blog posts. AI-shaped search rewards consolidation and clarity.
Why this is also an SEO/AEO/GEO play
When your content is structured and explicit, it’s easier for AI systems to cite, summarize, and match your brand to relevant prompts. This isn’t just “SEO for rankings.” It’s visibility inside answers—often described as AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization).
AYSA has a dedicated overview of this concept here: AI search visibility.
If you sell products: feed quality and Merchant Center fields become conversion levers
For ecommerce, AI-influenced search compresses the gap between “information content” and “product content.” Product data needs to carry more conversational context.
The SEJ reporting highlights that Google has encouraged merchants to fill additional fields in Merchant Center aimed at conversational experiences (again: treat as directional guidance; validate in your own account) (SEJ coverage).
Even without betting on any one field, the principle is sound:
- AI experiences prioritize “fit details” that help people decide.
- Merchant feeds that only include bare minimum attributes are less competitive when users ask nuanced questions.
Practical feed improvements that usually help (without guessing platform internals)
- Variant clarity (sizes, colors, compatibility)
- Accurate grouping (so comparisons make sense)
- Real-world Q&A surfaced consistently (on PDPs and, where possible, in feed enhancements)
- Policy clarity (shipping, returns, warranties)
The operational win is not “please the algorithm.” It’s reduce user uncertainty. AI search simply makes uncertainty more visible because users ask it out loud.
A concrete SME scenario: “The clinic that bought curiosity”
Let’s make this real with a scenario I see often.
The business
A regional clinic offers elective procedures (or specialized services like dermatology, fertility, physical therapy, dental implants). Their PPC historically focuses on bottom-funnel queries:
- “dermatologist near me”
- “laser treatment cost”
- “book dermatologist appointment”
Then AI-shaped search expands the query universe:
- “how do I know if I need to see a dermatologist for recurring rash”
- “best treatment options for acne scars sensitive skin”
- “what to expect after [procedure] recovery timeline and side effects”
What goes wrong if the clinic runs this like old PPC
- They broaden targeting to “catch more volume.”
- Clicks increase, but appointment bookings don’t move proportionally.
- CPA spikes, ROAS drops, and the owner declares PPC “isn’t working anymore.”
What a better approach looks like
Step 1: isolate discovery spend. Create a dedicated campaign group with a capped budget and a different primary KPI (engaged sessions to educational pages, consultation starts, email captures).
Step 2: build “decision support” landing pages. Not blog fluff—clinically reviewed pages that cover eligibility, contraindications, recovery, pricing ranges, and a clear “talk to a clinician” CTA.
Step 3: use remarketing ethically and helpfully. If someone reads the recovery timeline page, offer a checklist or a consult—not an aggressive discount ad.
Step 4: report assisted outcomes. Show how many booked appointments had an earlier visit to the educational hub within 30–60 days.
This is how you turn “curiosity clicks” into future revenue—and make that case to a business owner who doesn’t care about match types but cares about profit.
What agencies should rethink for 2027: deliverables, dashboards, and accountability
Agencies are going to feel this shift sharply because many client relationships are still anchored to a simple promise: “We buy high-intent search and deliver measurable leads/sales at target CPA.”
That promise becomes harder when the search landscape itself becomes more exploratory.
Three uncomfortable realities agencies need to address
- More of the work is on-site. If landing pages aren’t built for research behavior, paid performance will look worse than it “should.” Agencies that refuse to touch the website will lose.
- Attribution debates will increase. Clients will demand clarity exactly when clarity is hardest.
- Platform opacity is not your friend. If the ad platform doesn’t break out AI placement performance cleanly, agencies must build proxy measurement and experiments.
What the next-gen PPC retainer should include
- Query segmentation and intent mapping (not just keyword lists)
- Landing page briefs and conversion path optimization
- Cross-channel measurement (SEO + PPC + email + direct/brand)
- Experiment design and stakeholder reporting that explains uncertainty honestly
Deliverables that look like 2019 won’t survive 2027.
Where AYSA.ai fits: Approved execution for a messy, cross-channel reality
AI-driven search changes don’t fail because marketers lack ideas. They fail because execution is slow, fragmented, and hard to approve.
In most SMEs and even many mid-market companies:
- PPC sees new query patterns first.
- SEO owns content structure and internal linking.
- CRO owns forms and conversion friction.
- Product or legal owns policy language.
- Engineering owns release cycles.
So you get meetings, tickets, and delays—while the SERP changes weekly.
AYSA is built for this reality: it monitors your search visibility, prepares specific website changes, asks for approval, and then executes accepted changes. That model is especially relevant when AI search pushes you to iterate on pages, FAQs, comparisons, and entity-level clarity.
Where this connects directly to the AI Mode / upper-funnel shift:
- Monitoring: Track which pages gain/lose visibility and which topics become “question-shaped.” See: AYSA monitoring.
- AI search visibility: Optimize not just for rankings, but for being understandable and cite-worthy in AI answers. See: AI search visibility.
- Execution at scale: Turn PPC insight (“people keep asking X”) into site improvements (FAQ blocks, comparison sections, clearer policies, internal links) without waiting for a quarterly web sprint. Explore: AI SEO tools.
If you’re evaluating tools, you can review options here: AYSA pricing. And for additional playbooks, the ongoing library is here: AYSA blog.
My opinion: in AI-shaped search, “strategy” is table stakes. The compounding advantage is fast, controlled execution with a paper trail of approvals.
What to do next (action list)
If you want a practical sequence that doesn’t require rewriting your entire marketing program, start here.
Week 1: Diagnose the shift
- Export search terms and segment by word count (1–3, 4–7, 8+).
- Identify the top 20 long-query themes (comparisons, “best for,” “how to choose,” “for [use case]”).
- Check engagement metrics on the landing pages those queries hit (bounce rate alone isn’t enough; look for meaningful interactions).
Week 2: Isolate spend and set a test contract
- Create a dedicated “discovery” campaign group with a capped budget.
- Define success as a mix of: assisted conversions, engaged sessions, consultation starts, email captures, or product-detail depth.
- Build a one-page internal memo explaining the hypothesis and measurement approach.
Weeks 3–4: Fix the post-click experience
- Choose 3–5 landing pages that receive the most discovery traffic.
- Add “follow-up question” sections: comparisons, constraints, pricing clarity, policies, proof.
- Ensure internal links guide users from research → evaluation → conversion pages.
Month 2: Prove (or disprove) incrementality
- Run an experiment or holdout if feasible (geo split, time-boxed holdout, or campaign experiment).
- Report results as: what we learned, what we changed, what we’ll test next.
Ongoing: Build a system, not a one-off fix
- Set monitoring so you’re alerted when visibility shifts or new question themes emerge.
- Create an execution loop: propose changes, approve, implement, measure.
- Document decisions for finance and leadership (durable proof beats slides full of opinions).
Sources and further reading
- Search Engine Journal (primary reference): Paid Search Is Moving Up The Funnel – Google Admits AI Mode Pays Less
- Search Engine Journal topic hub (context): AI Search on SEJ
- Search Engine Journal topic hub (context): Paid Media on SEJ
- AYSA.ai: AI search visibility (AEO/GEO)
- AYSA.ai: Monitoring
- AYSA.ai: AI SEO tools
- AYSA.ai: Pricing
- AYSA.ai: Blog
Note on sourcing: The supplied research context references Google statements, a Google podcast discussion, and Alphabet earnings commentary as reported by Search Engine Journal. Where an official primary link (e.g., a help document, earnings transcript, or Google post) is not directly provided in the supplied context, I’ve treated details as directional and recommended validating within your own accounts and reporting.
Continue the AI search topic inside AYSA.
Use these pages to connect the article with AI SEO tools, AI visibility monitoring, AI Overviews and approved website execution.
Turn this topic into a website action plan.
Use these AYSA hubs to move from reading to technical fixes, AI visibility monitoring, research, glossary context and approval-first SEO execution.
