The quiet crisis: the internet is keeping your users
Across those headlines there’s one pattern that actually matters to operators right now:
Distribution is being eaten by platforms and AI surfaces that don’t send traffic back.
AI Overviews, answer engines, sponsored shops in SERPs, TikTok all-in-one funnels, LinkedIn as the top-cited source in AI search, declining referral traffic for smaller publishers – it’s all the same story:
- Search results that answer in-line instead of sending clicks
- Social platforms that keep the entire funnel on-platform
- AI assistants that summarize your work without attribution or a visit
For CMOs, performance marketers, and media buyers, this isn’t a thought experiment. It’s a P&L problem. Your old assumption – “do good content + buy smart media = get traffic to owned properties” – is quietly breaking.
This piece is about operating in a world where “clicks” are a lagging indicator and “visibility inside other people’s interfaces” is the leading one.
The new distribution stack: where attention actually lives
Look at a few of the headlines you shared:
- Google Is Testing Sponsored Shops in SERPs
- AI Overview Click Data Reveals Unexpected User Behavior Patterns
- Why LinkedIn Is the Most-Cited Source in AI Search
- From Content to Conversion: TikTok’s New All-in-One Funnel Tools
- Referral Traffic Is Declining for Smaller Publishers
- How to get indexed by ChatGPT [2026]
- FAQs for AEO: How to structure answers that rank in answer engines
These aren’t random product updates. They’re a map of the new distribution stack:
- Search results pages as full shopping and answer experiences (Sponsored Shops, AI Overviews, answer engines).
- Social platforms as end-to-end funnels (TikTok funnel tools, Facebook Shops, LinkedIn content → DM → call flows).
- AI assistants and answer engines as discovery layers (ChatGPT indexing, Claude visibility depending on Brave, LinkedIn as a primary citation source).
In all three, the platform’s goal is the same: keep the user, not send them to you.
If your operating model is still “rank → click → convert”, you’re playing last decade’s game with this decade’s budget.
Stop worshipping sessions; start managing surfaces
Most teams are still organized around channels:
- SEO team: “organic traffic”
- Paid search team: “clicks and ROAS”
- Social team: “engagement and followers”
- Brand team: “awareness”
The internet, meanwhile, is organized around surfaces – specific UI contexts where your brand can appear or disappear:
- Google AI Overview box for your category query
- Shopping units and Sponsored Shops carousels
- Reddit threads and Q&A snippets that feed AI answers
- TikTok product cards, live shopping, and in-feed checkout
- LinkedIn posts that get scraped into AI summaries
- ChatGPT / Claude / Perplexity answer panels
Each surface has its own:
- Ranking logic
- Attribution blind spots
- Creative constraints
- Commercial potential (even without a click)
The job now is to treat these surfaces as inventory you can win, not accidents you hope for.
A practical framework: measure, defend, expand
Here’s a simple operating framework you can actually run inside a growth or marketing org: Measure → Defend → Expand.
1. Measure: build “off-site impact” into your reporting
If your dashboards only show what happens on your properties, you’re flying blind. You need to see the value you create where the user never clicks through.
At minimum, add three layers:
a) Surface-level visibility metrics
For each key surface, track visibility as a first-class metric:
- Search & AI Overviews
- Share of voice on page-1 for core commercial queries (classic SEO)
- Presence in AI Overviews / answer boxes (manual tracking + SEO tools that now monitor this)
- Presence in “People also ask” and FAQ-style snippets
- Social commerce & funnels
- Share of category impressions inside TikTok shopping / Facebook Shops
- Click-to-message or click-to-shop rates from native tools
- DM volume and quality as a funnel entry metric, not just site visits
- AI assistants / answer engines
- Brand and product mention tests in ChatGPT, Claude, Perplexity for priority queries
- Frequency and quality of citations (are you a primary source or background noise?)
b) “Dark influence” proxies
You won’t get perfect attribution. You can still get directional signal:
- Correlate changes in AI / SERP visibility with:
- Brand search volume
- Direct traffic
- Category conversion rates on paid media
- Track “where did you hear about us?” in:
- Post-purchase surveys
- Sales calls
- Onboarding flows
- Use simple tags like “Google answer”, “ChatGPT”, “TikTok video”, “Reddit thread” instead of a 20-option survey that no one fills out accurately.
c) Unit economics that don’t assume a click
Redefine success for some surfaces as “qualified exposure” instead of “sessions”. For example:
- If your product is fully described and recommended in an AI Overview, that exposure has value even with no click.
- If TikTok’s funnel tools let a user view, consider, and buy without touching your site, your CAC math should treat that as a complete funnel, not “lost attribution”.
That means building channel P&Ls that accept partial data and using incrementality tests instead of obsessing over last-click.
2. Defend: stop losing the surfaces you already earned
While everyone chases the next AI toy, many brands are quietly losing existing surfaces through neglect.
Defensive moves that matter right now:
a) Fix cannibalization and fragmentation
Moz talking about cannibalization and 8,000 title tag rewrites is a symptom: too many brands are competing with themselves for the same surfaces.
- Consolidate overlapping content and product pages around clear intents:
- “What is X?” (education)
- “Best X for Y” (comparison)
- “X pricing” (commercial)
- Make one page the canonical authority for each important question. That’s what answer engines want to cite.
b) Structure your content for answer engines, not just humans
Headlines like “FAQs for AEO” and “How to get indexed by ChatGPT” are basically telling you: structure matters more than ever.
- Use tight Q&A formats for core questions (literally “Question:” / “Answer:” blocks).
- Write short, extractable definitions and summaries at the top of pages.
- Use schema markup for FAQs, products, and how-tos so machines can parse you cleanly.
- Keep your brand and product names close to the problem statement so AI models associate them.
c) Protect your brand from AI misrepresentation
“Using AI to Support and Defend Your Brand” and “AI’s trust problem” are the flip side: these models will happily hallucinate about you.
Minimum viable brand defense:
- Quarterly “AI brand audit”: ask major assistants:
- Who are we?
- What do we sell?
- Who are alternatives?
- What are our pros/cons?
- Where they’re wrong, fix your own surfaces:
- Clarify positioning and claims on your site.
- Clean up outdated third-party profiles and review sites.
- Publish clear, up-to-date comparison and “who we’re for / not for” content.
- For serious inaccuracies, use the platform’s feedback channels – but assume your own content footprint is the main training data you control.
3. Expand: design for “zero-click growth”
Once you’ve measured and defended, you can start playing offense: growing impact even when you don’t get the click.
a) Treat platforms as full-funnel, not just top-of-funnel
Headlines about TikTok’s all-in-one funnel tools and Facebook Shops aren’t just shiny features. They’re an invitation to stop forcing users onto your site when they don’t need to go there.
Practical moves:
- Stand up native product catalogs and keep them clean (inventory, pricing, imagery, reviews).
- Design platform-native offers (TikTok-only bundles, Instagram-only drops) that make sense to complete in-app.
- Use click-to-message and click-to-call as primary CTAs for high-consideration products instead of “learn more” site links.
b) Optimize for being the cited expert, not just the ranked result
Why is LinkedIn the most-cited source in AI search? Because the content is:
- Topical and fresh
- Authored by identifiable experts
- Structured into clear, quotable chunks
Translate that into your playbook:
- Invest in named experts (founders, PMs, lead practitioners) publishing under their own names on LinkedIn, industry blogs, and Q&A communities.
- Focus them on narrow, repeated topics you want to own in AI answers, not random thought leadership.
- Repurpose those posts into on-site content that reinforces the same messages and gives answer engines a clean source to cite.
c) Design creative for “seen once, remembered later”
If you accept that many exposures won’t click, your creative brief changes. The question becomes: “If someone only sees this once, with no click, what sticks?”
Adjust your ads and content to:
- Make the category + brand + promise obvious in three seconds:
- “Payroll for 2-50 person agencies – in 5 clicks.”
- “The ecommerce ESP that fixes broken flows automatically.”
- Use distinctive brand assets that survive in tiny placements:
- Color + shape + short name beat clever headlines in cramped surfaces.
- Assume muted, half-watched, or screenshot consumption. Design for that.
How to reorganize your team around this reality
This isn’t just a tactics shift; it’s an org design problem. If your teams are siloed by channel, no one owns the surfaces that sit between them.
Three structural changes that help:
1. Create a “Search & Surfaces” function
Merge traditional SEO, parts of content, and some analytics into a single team responsible for:
- Visibility across SERPs, AI Overviews, answer engines, and marketplaces
- Content architecture and schema for machine readability
- Surface-level reporting and experimentation
Give them a mandate that explicitly includes off-site impact, not just organic traffic.
2. Give performance marketing permission to optimize for assisted impact
If your media buyers are punished for anything that doesn’t show up as last-click ROAS, they’ll ignore high-impact surfaces that don’t send traffic cleanly.
Change the brief:
- Include incremental lift tests in their scorecard (geo splits, holdout audiences, time-based tests).
- Reward category share-of-voice gains on key surfaces (shopping units, in-feed placements) even when attribution is messy.
- Budget a fixed percentage (say 10-15%) for “zero-click” experiments with clear hypotheses.
3. Put brand and performance on the same calendar
Right now, brand is often making “big moments” while performance chases daily efficiency. In a world of AI summaries and on-platform funnels, consistency beats stunts.
Run a shared calendar where:
- Every major campaign has:
- On-site assets
- Platform-native executions
- Structured content for answer engines
- Brand guardrails include:
- How we want to be described in one sentence
- Which problems and categories we want to be associated with
- Performance teams pressure-test whether that actually shows up in search queries, AI answers, and social comments.
What to do in the next 90 days
If you’re leading a team, here’s a concrete 90-day roadmap:
- Audit your surfaces
- List your top 20-30 revenue-driving queries, categories, and problems.
- Check how you show up across: Google (including AI Overviews), TikTok, Instagram, Facebook Shops, LinkedIn, Reddit, and at least one AI assistant.
- Score each surface: “Absent / Present but weak / Present and strong”.
- Patch the obvious holes
- Fix missing or broken product feeds for shopping and social commerce.
- Publish or clean up core Q&A and definition content on your site.
- Resolve the worst cannibalization issues where you compete with yourself.
- Ship one “zero-click” experiment per major channel
- Search: optimize explicitly for an AI Overview answer on one high-intent query.
- Social: run a TikTok or Instagram funnel that completes entirely in-app.
- AI: design one authoritative resource and test how it changes how assistants describe your brand over time.
- Update your dashboards
- Add surface visibility metrics and simple “where did you hear about us?” data.
- Start reviewing these in your weekly growth or marketing meeting.
The platforms and AI layers are not going to start sending more traffic back. Their incentives are clear. Your job is to treat that as a design constraint, not a tragedy – and build a growth engine that assumes the click is a bonus, not the baseline.