The pattern nobody wants to admit: you’re still optimizing for 2018 Google
Scan those headlines and a single theme jumps out: search is quietly mutating from “10 blue links” into an answer layer run by AI agents – and most marketing teams are still behaving like that’s a side quest, not the main game.
We’ve got:
- AI Overviews and “answer engines” eating clicks from both SEO and paid search.
- Articles on how AI agents “see” your site, and how Google’s own leaders are openly describing a more AI-native search experience.
- CMOs adding AI to their remit, while teams debate if AI content is “bad for SEO” instead of asking a better question: what is SEO even optimizing for now?
The real issue: most orgs are still running a “rank and bid” playbook in a world that’s moving to “answer and orchestrate.”
If you’re a CMO, performance lead, or media buyer, the risk isn’t that AI steals your job. It’s that you keep measuring and optimizing for a search reality that no longer exists – and find out 12 months too late.
From search engine to answer engine: what actually changed
Forget the hype. Practically, three things now matter:
1. The click is no longer the default outcome
AI Overviews, rich snippets, “People also ask,” and soon more agent-like interfaces mean:
- More queries end on the SERP with no click.
- When a click does happen, it’s often later in the journey and closer to intent.
- Paid units are increasingly blended into “answers” instead of clean ad blocks.
Your old mental model – impression → click → site → conversion – is now impression → answer → maybe click → conversion. That “answer” step is not neutral; it’s a powerful filter.
2. AI agents are new gatekeepers to your content
Tools and agents (including those from Google, OpenAI, Anthropic, and others) are:
- Parsing your site differently than humans do.
- Prioritizing clarity, structure, and explicitness over clever design.
- Summarizing and recombining your content with competitors’ into synthetic answers.
If your site is a maze of modals, JS-heavy experiences, and vague copy, humans might still muddle through. Agents won’t. They’ll skip you.
3. Context beats content volume
The arms race to publish more content with AI is missing the point. In an answer-driven world:
- Who you are (brand, authority, expertise) matters more than how much you publish.
- Where your content sits in the customer journey matters more than raw traffic.
- Signals of real-world credibility (reviews, PR, expert authors, user proof) are increasingly important inputs to AI systems.
You don’t need 10x more content. You need 10x clearer context around what you do, for whom, and why you’re the safest answer.
The uncomfortable measurement gap
High-growth companies are rethinking how they measure marketing. Most others are still arguing about “last click vs data-driven” while their actual problem is simpler: they’re measuring the wrong game.
Three gaps show up repeatedly in operator conversations:
Gap 1: Reporting on clicks while losing share of answers
Dashboards still show:
- Impressions
- Clicks
- CTR
- ROAS / CPA
What’s missing:
- How often you appear in AI Overviews or answer modules.
- Whether your brand is even mentioned in synthetic answers for your core queries.
- Share of “answer real estate” versus competitors on high-intent topics.
You can be “stable” on branded search CTR and still be losing the unbranded answer layer that feeds your future pipeline.
Gap 2: Misreported ROAS in an automated, AI-heavy ad stack
With auto-bidding, broad match, and black-box optimizations:
- ROAS is increasingly a model output, not a ground truth.
- Attribution windows and conversion modeling hide demand that never clicks.
- Brand and performance bleed into each other, but budgets and KPIs stay siloed.
The result: teams over-invest in what’s easy to measure (bottom-funnel branded, retargeting) while under-investing in the content and context that make AI engines choose them upstream.
Gap 3: Content metrics that ignore how AI tools use your site
Content teams still celebrate:
- Organic traffic growth.
- Average position for target keywords.
- Time on page.
But AI agents don’t care how engaging your narrative is if they can’t:
- Quickly understand what you do.
- Extract clean, structured answers.
- Map your content to specific intents or use cases.
You’re optimizing for human dwell time while the new gatekeepers are optimizing for machine readability.
What operators should actually do in the next 12-18 months
You don’t need a 50-slide “AI transformation” deck. You need a focused operating plan that acknowledges the answer engine reality and moves budget, content, and measurement accordingly.
1. Redesign your search strategy around “answer share,” not keyword rank
Keep doing keyword research – but change the output:
- Group queries into answer themes, not just keyword clusters. Example: “pricing”, “implementation”, “alternatives”, “ROI”, “security”, “industry-specific use cases.”
- For each theme, map:
- Which SERP features show (AI Overviews, FAQs, videos, local, shopping).
- Which competitors or publishers consistently appear in those features.
- Where your brand appears today – if at all.
Then set a simple north star for search:
“For our top 20 answer themes, we will be mentioned or featured in at least 60% of AI/answer surfaces within 12 months.”
That’s what you brief content, PR, and paid teams against – not “rank #1 for [category] software.”
2. Make your site legible to AI agents
This is not about chasing every schema type. It’s about making your website a clean knowledge base for both humans and machines.
Run a ruthless audit with three questions:
- Can an agent understand what we do in 10 seconds?
- Plain-language positioning above the fold.
- Clear product taxonomy and use cases.
- No jargon salad on your homepage hero.
- Are our core answers explicit and structured?
- Dedicated pages or sections for pricing, implementation, security, integrations, ROI.
- FAQ blocks with concise Q&A, not fluffy paragraphs.
- Tables, comparison charts, and bullet lists that are easy to parse.
- Is our expertise machine-visible?
- Author bylines with real credentials.
- Case studies with specific numbers and named customers where possible.
- Structured data for organization, products, reviews, and FAQs.
You’re not decorating your site for Google. You’re making it trivial for any AI system to extract “who you are, what you do, why you’re credible.”
3. Treat AI content as a drafting tool, not a strategy
The “is AI content bad for SEO?” question is a distraction. The better question:
“What parts of our content deserve human judgment, and what parts can be automated without hurting trust?”
A practical split:
- Automate: outlines, first drafts of commodity pages, meta descriptions, ad variants, simple FAQs, localization drafts.
- Human-own: narratives that define your category, opinionated thought leadership, detailed case studies, complex product education, anything that carries brand risk.
Then enforce a non-negotiable rule: no AI output goes live without a clear point of view and a named owner. AI can synthesize; only your team can decide what you actually want to say.
4. Rebuild paid search around intent bands, not endless SKUs
The days of hyper-granular account structures are fading, but that doesn’t mean “let the algorithm do everything.”
Instead, structure your accounts by intent bands that match the answer themes you mapped:
- Problem-aware (symptoms, “how to fix X”).
- Solution-aware (categories, “best [type] tools”).
- Brand-aware (your brand, competitors, alternatives).
- Decision-ready (pricing, demo, free trial, near-me, coupons).
For each band:
- Define the business outcome (demo, trial, add to cart, store visit).
- Set guardrails for automation (bidding strategies, negatives, budgets).
- Align landing pages and on-site answers to the intent – not generic homepages.
Then measure ROAS and CAC by intent band, not just at account level. This is where you’ll see the real impact of AI Overviews and answer features on performance.
5. Fix your ROAS story before finance does it for you
In an AI-heavy environment, misreported ROAS isn’t a rounding error; it’s how budgets die.
A simple, defensible approach:
- Separate “measurable performance” from “answer presence.”
- Performance: channels and campaigns with clear, attributable conversions.
- Answer presence: content, PR, and search initiatives that increase your share of answers and assisted conversions.
- Use mixed models, not just platform numbers.
- Run regular incrementality tests (geo splits, holdouts) on branded and non-branded search.
- Use media mix modeling or lightweight regression to understand the impact of “answer presence” on direct and brand traffic.
- Report in language the CFO understands.
- Translate answer share and AI presence into leading indicators: lower CAC over time, higher close rates, shorter sales cycles.
- Show how cuts to “unmeasured” answer-building work degrade those indicators with a 3-6 month lag.
Your job is to prevent “AI ate our clicks” from becoming “so we cut everything that wasn’t last-click profitable.”
6. Put someone in charge of AI context, not just AI tools
Many CMOs now “own AI,” which often means they own a stack of tools. That’s not enough.
You need a clear owner for AI context:
- How your brand is represented in AI answers across major platforms.
- What internal knowledge (docs, FAQs, product specs, case studies) is available to your own AI systems and agents.
- How your messaging and positioning show up in synthetic content generated by others.
Give that person a mandate and budget to:
- Audit external AI outputs for your category and brand monthly.
- Work with product, sales, and support to keep your knowledge base current and structured.
- Set guidelines for how your teams use AI in copy, creative, and customer comms.
The real AI race isn’t who has the flashiest tool stack. It’s who controls the context that those tools draw from.
What to stop doing this quarter
If you want a quick sanity check, here are three things to cut or change now:
- Stop shipping content with no obvious “answer job.” If a page doesn’t clearly answer a real user question or move someone to a defined next step, it’s noise – for both humans and AI.
- Stop treating SEO, paid search, and content as separate sports. In an answer engine world, these are just different ways of influencing the same outcome: when someone asks a question, do we show up as the safest choice?
- Stop accepting platform-reported ROAS at face value. Use it as a directional signal, not a single source of truth. Layer on incrementality tests and intent-band views.
The operators who win the next few years won’t be the ones who publish the most AI content or chase every new feature. They’ll be the ones who quietly rebuild their marketing around a simple idea:
when someone asks a question – to Google, to ChatGPT, to an in-product assistant – our brand is the most obvious, credible, and easy-to-choose answer.