Search just split in two – and most teams are only budgeting for one side
Look at those headlines and a pattern jumps out:
- Google Search sends only 23% of queries to the open web.
- Claude is suddenly a meaningful traffic source.
- Google is testing Sponsored Shops in SERPs.
- Answer Engine Optimization (AEO), schema for AEO, and FAQ structuring are now mainstream topics.
- LinkedIn is the most-cited source in AI search.
- Google can be held liable for false AI Overview claims.
Translation: “search” is no longer just ten blue links and a few ad slots. It’s a duopoly:
- The classic web SERP (organic + paid).
- AI answer surfaces (AI Overviews, Claude, ChatGPT, Perplexity, answer boxes, AEO).
Most marketing orgs are still operating as if the second one is a side quest. It’s not. It’s where the next wave of intent is going.
If you own a performance budget or a P&L, you can’t afford to treat AI answers as “SEO’s problem” or “something content will look at later.” You need a channel strategy, a measurement strategy, and a budget strategy for AI-native search.
What’s actually changing (beyond the hype)
1. Discovery is shifting from “click” to “answer”
The open web is now the fallback for many queries, not the starting point. Headlines about:
- Google Search sending only 23% of queries to the open web
- Declining referral traffic for smaller publishers
- AI Overviews and answer engines
all point to the same thing: users increasingly get what they need from a synthesized answer, not your page.
That doesn’t mean your brand disappears. It means:
- Your content is training data for someone else’s interface.
- Your brand may be cited (or not) even when you “win” the query.
- Your traffic curve decouples from your influence curve.
2. AI search has its own “ad products” already forming
Google testing Sponsored Shops in SERPs is not a one-off. It’s the first visible step in a broader pattern:
- AI answer surfaces will get monetized.
- Those monetization units will favor structured, high-integrity data.
- They’ll be sold as performance media, not just “brand exposure.”
The “future of TV outcome-based buying” and “LiveRamp as ChatGPT’s first conversion API partner” are the same story in different channels:
platforms are racing to tie AI-driven attention to measurable outcomes.
3. AI search has a trust and liability layer
When a German court says Google can be directly liable for false AI Overview claims, a new constraint appears:
- Platforms need verifiable, high-integrity sources.
- They need to show where answers come from (citations, graphs, integrity layers).
- They will favor entities they can “trust” programmatically.
That’s why you’re seeing concepts like the “Integrity Graph” and why LinkedIn is the most-cited source in AI search:
platforms like sources that are structured, identity-anchored, and relatively low-risk.
The operator problem: your stack is built for the wrong search
Most orgs still treat search as:
- SEO = rank pages in Google.
- SEM = buy keywords in Google Ads.
- Reporting = sessions, CTR, CPC, ROAS.
That stack fails in an AI-dual world because:
- Rank is less important than inclusion in the answer.
- Clicks are less important than citations, mentions, and assisted conversions.
- Keyword-level bidding is only half the game; you also need to influence model-level behavior.
The risk is simple: you keep optimizing for a shrinking surface (classic SERPs) while competitors quietly dominate the interfaces where users actually get their answers.
Treat AI answers as a channel: a practical blueprint
This is not about chasing every shiny AI product. It’s about treating AI answer surfaces as a real channel with:
- Inputs you can control.
- Signals you can measure.
- Budget you consciously allocate (or don’t).
Step 1: Map your “answer surface” footprint
Start with a simple audit across three layers.
Layer 1: Classic SERP features
- Where do you appear in:
- Featured snippets
- People Also Ask
- FAQ rich results
- Local packs
- Which of your high-value queries now trigger:
- AI Overviews
- Shopping units (e.g., Sponsored Shops)
- Zero-click answers
Have your SEO team pull a list of your top 200-500 revenue-driving queries and mark:
- AI Overview present? (Y/N)
- Shopping/commerce unit present? (Y/N)
- Are you cited in the AI Overview? (Y/N)
- Are you present in any rich result? (Y/N)
Layer 2: AI-native search (chatbots and answer engines)
For your category and brand terms, test:
- ChatGPT / GPT-based engines
- Claude
- Perplexity
- Any vertical AI tools your buyers use (coding, marketing, research, etc.)
Ask questions your buyers ask:
- “Best for [use case]”
- “[Problem] for [segment]”
- “Compare [your brand] vs [competitor]”
Track:
- Are you mentioned at all?
- Are your competitors mentioned more often?
- Are any of your owned properties cited (site, docs, LinkedIn, YouTube)?
Layer 3: Owned “answer assets”
Inventory:
- FAQ pages and help centers.
- Schema markup (FAQ, HowTo, Product, Organization, Person).
- Public documentation, whitepapers, and comparison pages.
- Executive and brand presence on LinkedIn (and other identity-rich platforms).
The goal: understand how much structured, machine-readable, identity-anchored content you actually have. That’s what feeds AI answers.
Step 2: Build an “AI visibility” KPI stack
You can’t manage what you don’t measure. The good news: you don’t need perfect tracking to start. You need directionally useful metrics.
Core KPIs to add to your dashboards
- AI answer presence rate: % of priority queries where:
- An AI Overview appears, and
- Your brand is cited or mentioned in that overview.
- Answer engine brand share: in AI tools, for key category queries, how often you’re mentioned vs. your top 3 competitors.
- Cited-source mix: what domains are being cited when AI tools answer questions in your category (e.g., LinkedIn, G2, niche blogs, your site).
- AI-assisted conversions: using prompt tracking or self-reported attribution (“How did you hear about us?” with AI tools as explicit options).
This is where the “prompt tracking” conversation becomes practical. You don’t need a perfect log of every prompt. You need:
- Qualitative feedback from sales calls and chat logs (“I asked ChatGPT about…”).
- Simple fields in lead forms and post-purchase surveys that include “AI assistant / AI search” as a channel.
- Periodic audits of AI tools to see how your share of mentions changes over time.
Step 3: Design content for answer engines, not just crawlers
A lot of “AEO” advice is just SEO with a new label. You need to go one level deeper: design content that is:
- Atomic: clear Q&A pairs that can be lifted into an answer.
- Structured: schema markup that tells machines what’s what.
- Attributable: strongly tied to your brand and people.
Practical moves
- Turn your best-performing pages into FAQ clusters with explicit questions and concise answers at the top.
- Add FAQ and HowTo schema where it actually matches the content (don’t spam).
- Publish clear comparison pages (“[You] vs [Competitor]”) with factual, sourced claims.
- Make your executives and experts visible on LinkedIn with consistent, deep content on your category – AI tools love citing named experts from identity-rich platforms.
- Ensure your robots.txt and meta settings aren’t accidentally blocking AI crawlers you actually want (while still protecting sensitive areas).
You’re not just chasing rankings; you’re feeding structured, attributable facts into the systems that now answer your buyers’ questions.
Step 4: Treat AI answer surfaces like a paid channel in planning
You can’t buy “top spot in AI Overview” (yet), but you can and should treat AI visibility as a budgeted initiative:
- Allocate owner time: someone senior in search or growth owns “AI visibility” as a specific responsibility.
- Fund experimentation: small budget for tools (prompt tracking, AI search monitoring, schema automation).
- Integrate into media planning: when you plan SEM/paid social, ask:
- Which queries are AI-heavy and deserve more content investment vs. more paid spend?
- Where will Sponsored Shops or new commerce units cannibalize organic – and should we buy in?
For commerce-heavy categories, assume:
- AI answers will handle discovery and education.
- Sponsored units (like Shops) will handle transactional intent.
Plan your budget accordingly instead of pretending organic will carry both.
Step 5: Build your own “integrity graph”
If AI tools and search engines are building integrity graphs to decide who to trust, you should be doing the same internally:
- Maintain a source of truth library for:
- Product specs
- Pricing logic
- Security and compliance details
- Customer proof (case studies, reviews, testimonials)
- Make that library:
- Public where possible (docs, knowledge bases, technical blogs).
- Consistent across channels (site, LinkedIn, sales decks, PR).
- Tag and structure it so internal and external AI tools can consume it cleanly.
This is not just a content exercise; it’s a risk and brand exercise. If Google can be liable for false AI claims, so can you, indirectly, when AI tools hallucinate about your product. Clear, consistent, structured information reduces that risk and improves your odds of being the cited source.
What CMOs and performance leaders should do in the next 90 days
If you want something concrete to act on, here’s a 90-day plan that fits into existing workflows.
Weeks 1-3: Baseline and ownership
- Appoint an AI visibility owner (usually head of SEO or growth) with a clear mandate.
- Run the answer surface audit on your top 200-500 revenue-driving queries.
- Test Claude, ChatGPT, Perplexity, and others for 20-30 high-intent queries and document brand/competitor mentions.
- Add “AI assistant / AI search” as an option in lead forms and post-purchase surveys.
Weeks 4-8: Content and structure
- Turn your top 10-20 money pages into answer-optimized assets:
- Clear questions and concise answers at the top.
- FAQ and HowTo schema where relevant.
- Updated, factual, sourced claims.
- Publish or update 3-5 comparison pages (you vs. key competitors, or solution alternatives).
- Launch a LinkedIn content cadence for at least one credible executive focused on your category’s core questions.
Weeks 9-12: Measurement and media integration
- Add AI answer presence rate and answer engine brand share to your main marketing dashboard.
- Review your next SEM and paid social plan with AI surfaces in mind:
- Shift some content budget toward queries where AI Overviews dominate.
- Test any new commerce units (e.g., Sponsored Shops) where they’re cannibalizing your organic.
- Run a sales and CS enablement session so frontline teams know how to talk about “I saw X in ChatGPT/Claude/Google AI Overview.”
The operators who win the next phase of search won’t be the ones who obsess over every new AI feature announcement. They’ll be the ones who quietly treat AI answers as a real channel: modeled, measured, budgeted, and managed with the same discipline as paid search and social.