The real story behind all those “Is SEO dead?” headlines
Ignore the drama. What’s actually happening is simpler and more brutal:
AI systems are reallocating attention inside platforms, and your media mix hasn’t caught up.
Look at the pattern in the headlines:
- “AI Overviews Reduce Clicks by 58%” and “How to Track AI Overviews”
- “How to structure pages for AEO and answer engines”
- “Information Retrieval: How To Get Into Model Training Data”
- “AI-Powered Functionality in Google’s SEO Tools”
- “PPC Skills That Won’t Be Replaced By Automation”
- “Future of TV: AI agents prime TV advertising for ‘premium automation’”
Across search, social, TV, CRM, and creative, the same thing is happening:
AI is absorbing the “top of funnel” experience and compressing the journey. Platforms are
keeping more of the interaction on their own turf. Your old playbook assumed a click. The new one can’t.
This isn’t an SEO story. It’s a traffic reallocation story. And it’s already a P&L problem.
From “optimize for clicks” to “optimize for answers”
For the last 15 years, the operating assumption of digital marketing has been:
get the click, then do your work on your own property.
AI overviews, answer engines, and social feeds driven by recommendation models are changing that:
- Search results answer the question directly and may never send the user to your site.
- Social platforms summarize, remix, and keep users scrolling in-app.
- AI agents (in TV, CRM, and support) route people to outcomes without exposing your “owned” experiences.
The result: your historical metrics (sessions, CTR, last-click ROAS) are becoming
less representative of reality, even if your dashboards still look “fine.”
The operators who win the next three years will treat AI systems as
distribution channels and surfaces, not just tools.
The three big shifts CMOs and media buyers can’t ignore
1. Impression value is migrating from pages to snippets
When AI overviews cut clicks by 58%, two things happen:
- Your organic traffic drops, even if your “visibility” looks good.
- The unit of value moves from “visit to your site” to “inclusion in the answer.”
That’s the same pattern we’ve already lived through:
- Facebook: from page visits to in-feed consumption.
- Instagram/TikTok: from profile visits to content impressions.
- Now search: from website visits to answer snippets and citations.
If your team is still reporting “organic traffic” without segmenting:
brand vs non-brand, AI vs classic SERP, answer vs non-answer, you’re flying blind.
2. Optimization is drifting from manual tactics to problem deduction
Look at the skills articles:
- “The Real SEO Skill No One Teaches: Problem Deduction”
- “The PPC Skills That Won’t Be Replaced By Automation”
- “AI Isn’t Replacing Humans. It’s Reallocating Human Judgment.”
Platforms are automating the knobs you used to turn:
bids, placements, creative rotations, even keyword matching. What they don’t automate:
- Defining the actual business problem.
- Interpreting messy, partial data when the platform hides key metrics.
- Designing experiments that reveal what the black box is doing.
In other words: the scarce skill is no longer “how to edit 8,000 title tags.”
It’s “how to decide whether rewriting 8,000 title tags is even the right problem.”
3. “Owned media” is being redefined by model training data
You’re no longer just fighting for rankings. You’re fighting for
inclusion in model memory.
Articles like “How To Get Into Model Training Data” and “How to structure pages for answer engines”
are pointing at a deeper shift: the most valuable “placements” may be:
- Being a cited source in AI overviews.
- Being the canonical example in a model’s answer to “best X for Y.”
- Being the brand name that appears in generated recommendations and summaries.
That’s not classic SEO. It’s model-era brand distribution.
What this means for your media mix in practice
You don’t need a new buzzword. You need to change how you plan, buy, and measure.
Here’s a practical way to adapt without blowing up your org chart.
1. Add “answer surfaces” as a distinct line in your media plan
Today your plan probably looks like:
- Paid search
- Paid social
- Programmatic / CTV
- Organic search / content
- Email / CRM
You need a new line item: Answer & agent surfaces.
That includes:
- AI overviews and answer boxes in search.
- Chat-based experiences (on-site, in-app, or via partners).
- AI agents in TV, commerce, and CRM that mediate choices.
For each, define:
- Surfaces: where answers appear (SERP AI overview, in-app chat, TV agent, etc.).
- Roles: awareness, consideration, or direct response.
- Inputs: what content, feeds, or schema they draw from.
- Signals: what you can actually measure (citations, mentions, assisted conversions).
2. Move from “click-based” to “influence-based” measurement
If AI overviews and social feeds are doing more of the explaining,
then your attribution model has to recognise influence without a click.
At a minimum, add three layers to your reporting:
-
Visibility metrics
Track:- AI overview presence and citations for your brand and competitors.
- Share of answers for key non-brand queries.
- In-platform engagement where no click is required (saves, shares, replies, DMs).
-
Assisted impact
Tie visibility to:- Branded search volume over time.
- Direct traffic with no clear referrer.
- Lift in conversion rate among exposed cohorts (where you can model it).
-
Downstream conversion
Stop obsessing over last-click ROAS from channels that are increasingly top-of-funnel.
Use MMM or at least structured holdout tests to quantify their real contribution.
The key operational change: your weekly performance review should include
answer visibility and influence metrics, not just spend, clicks, and CPA.
3. Treat content as structured fuel for models, not just pages for humans
The “how to structure pages for answer engines” and “get into model training data” threads
point to a practical shift: content needs to be machine-legible first.
That doesn’t mean writing for robots. It means:
- Clear, unambiguous answers high on the page.
- Consistent entities: product names, categories, benefits, and use cases.
- Schema markup that actually reflects your offer and pricing.
- FAQs that mirror how people ask questions in natural language.
Operationally, this looks like:
- Adding a “model-readiness” checklist to every major content or landing page brief.
- Reviewing top non-brand queries and ensuring you have concise, quotable answers.
- Creating canonical “explainers” for your category and making them technically clean.
4. Re-skill your performance team around problem deduction
The PPC and SEO skills that survive automation are the ones that sit upstream:
defining problems, designing tests, and interpreting ambiguous outcomes.
As a CMO or head of growth, ask your team to start presenting work in this format:
- Problem statement: business problem, not channel problem.
- Hypothesis: what they believe is happening in the platform black box.
- Experiment design: what they’ll change, where, and how they’ll know if they’re right.
- Decision rule: what they’ll do with the result (scale, stop, or pivot).
Then measure them on:
- Speed and quality of problem framing.
- Number of high-signal experiments run per quarter.
- Impact of decisions made from those experiments.
This is how you turn “AI won’t replace judgment” from a slogan into an operating model.
Where to reallocate budget in the next 12 months
You don’t need a revolution. You need a rebalancing. Here’s a pragmatic way to shift spend.
1. Defend your brand terms and high-intent surfaces
With AI compressing journeys, more people will search your brand once and expect an answer.
Make sure:
- Your brand terms in search and marketplaces are fully defended.
- Your knowledge panels, Maps listings, and profiles are accurate and rich.
- Your site can actually convert the traffic you still get (CRO is now a force multiplier).
This is boring, unsexy work. It’s also where a lot of incremental profit will come from.
2. Invest in “answerable” content and formats
Shift some of your generic blog/content budget into:
- Category explainers designed to be quoted in AI answers.
- Short, high-signal videos that can be summarized and recommended by feeds.
- Structured guides and FAQs that map to real queries, not just keywords.
Hold this content to a different KPI: answer visibility and assisted conversions,
not just pageviews.
3. Pilot at least one AI-native distribution channel
Don’t wait for a perfect playbook. Run controlled pilots where AI is the primary interface:
- An AI-assisted on-site buying guide that you treat as a “micro-channel” with its own funnel.
- AI-powered CRM journeys that adapt messaging based on behavior, not static segments.
- Tests with TV or retail media networks that use AI agents for planning or optimization.
The goal isn’t vanity “AI usage.” It’s learning how these systems behave when they’re
the ones deciding what your customers see next.
The uncomfortable but useful mindset shift
For twenty years, digital marketing has been built on one fantasy:
that you can own the journey because you own the click.
The headlines are telling you that era is ending:
AI overviews, answer engines, automated media buying, AI CRM, AI TV agents.
They all point to the same reality: platforms and models now own more of the journey than you do.
Your job isn’t to fight that. Your job is to:
- Make sure your brand is present and persuasive in the answers themselves.
- Measure influence, not just traffic.
- Point human judgment at the right problems while the machines handle the grunt work.
The teams that adapt their media mix to this traffic reallocation will quietly compound an advantage
while everyone else argues about whether SEO is “dead.” It’s not dead. It’s just moved.