The quiet collapse of the click
Look at those headlines again and a pattern jumps out: AI overviews, answer engines, “Read more” links, AEO vs. SEO, Facebook’s new rules for reach, consumers buying directly on social, WhatsApp as a channel, utility content for AI search.
Translation: the old “rank, get the click, convert on-site” funnel is being quietly dismantled in front of us.
AI overviews, answer engines, social commerce, and messaging-based journeys are all doing the same thing: compressing the funnel and stripping out the click. Google, Meta, Amazon, and TikTok are not “stealing traffic”; they’re finishing more of the journey on their own turf.
For CMOs, performance leaders, and media buyers, this is not a thought experiment. It’s a budget problem, an attribution problem, and a planning problem. If your operating model still assumes “traffic is the product,” you’re about to spend a lot of money optimizing for behavior that never happens.
The new reality: answers, not visits
Three shifts from those headlines that actually matter to operators:
- Answer engines are now a distribution layer. “Are AI overviews stealing your clicks?” and “Answer engine optimization case studies” are both saying the same thing: the SERP is no longer a list of doors; it’s a destination. Your content is fuel, not the final stop.
- Platforms are hoarding the journey. “Consumers are taking a new purchase journey on social,” “WhatsApp marketing,” “Facebook’s 2026 rules for reach,” “Read more” deep links – these are all about platforms keeping users in-app and compressing awareness, consideration, and conversion into a single environment.
- AI is a gatekeeper, not just a tool. “Why ChatGPT cites one page over another,” “What AI writing tools get wrong,” “Advanced AI deep research” – AI systems are deciding which brands even appear in the answer layer, and most marketers are still treating them like copy interns.
The net effect: your media and content are increasingly consumed without a click, a visit, or a pixel fire. If your KPIs and structures are built around sessions and last-click ROAS, you’re flying blind.
Stop asking “How do we get the click?” Start asking three better questions.
You can’t brute-force your way back to 2018. Instead, reframe around three questions:
- How do we show up in answer surfaces and in-platform journeys?
- How do we signal quality to AI systems and algorithms that decide what to surface?
- How do we measure business impact when the click is missing or delayed?
Let’s turn those into operating moves.
1. Design for “zero-click value” first, traffic second
Most brands still create content with one goal: get the click. That’s backwards now. You need assets that do their job even if the user never leaves the platform.
What “zero-click value” looks like in practice
- AI-overview-ready pages. Clear definitions, concise step lists, structured FAQs, clean headings, and schema markup. Think “if an answer engine scraped this page, would it have everything it needs to answer well and still attribute us?”
- Utility news and explainers. Search Engine Land’s “utility news content” framing is right: content that is updated, specific, and directly useful (calculators, timelines, checklists, decision trees) gets referenced more often than fluffy thought pieces.
- Social-native education. If “consumers are taking a new purchase journey on social,” your product education and objection handling need to live in Reels, Shorts, carousels, Stories, and DMs – not just gated PDFs and blog posts.
- Commerce-ready messaging. WhatsApp, Messenger, and Instagram DMs are no longer “support channels”; they are mid-funnel and bottom-funnel. Build flows where the entire journey can happen in a thread.
The tactical shift: brief content and creative teams on “answer first, click optional.” If a user gets everything they need from an AI overview, a carousel, or a DM thread and then buys later via branded search, that’s a win – even if your web analytics never saw the original touch.
2. Treat AI and algorithms as B2B audiences
You now have two audiences:
- The human who buys
- The machine that decides whether the human ever sees you
Most teams obsess over the first and ignore the second. That’s why we’re getting a wave of “Why ChatGPT cites one page over another” and “What AI writing tools get wrong” content. Machines are picky. They care about structure, clarity, and consistency more than cleverness.
How to send stronger signals to answer engines and feeds
- Authoritativeness over volume. Cannibalization and 8,000 title tag rewrites are symptoms of the same disease: thin, overlapping pages. Consolidate into fewer, deeper, clearly scoped assets that own a topic. Machines hate ambiguity.
- Explicit topical focus. If you want to be cited in AI answers about “B2B payments for marketplaces,” your site, content clusters, and off-site mentions should all reinforce that you are about B2B payments, not “fintech, SaaS, and innovation” in general.
- Structured data and consistent entities. Use schema, consistent naming, and clean internal linking. This is boring, unsexy work. It is also exactly the kind of thing answer engines reward.
- Signals beyond your site. Digital PR “duplication” methods and partnerships matter more when AI models are trained on the open web. Repeated, consistent mentions of your brand + topic across credible domains raise your odds of being the cited example.
- Machine-readable performance. Fast pages, accessible layouts (note the WCAG 2.1 AA push), stable mobile UX – not because “Google likes it,” but because models and ranking systems use these as proxies for quality.
The mental model: you’re doing B2B marketing to AI systems and feeds. They have selection criteria. Treat them as stakeholders, not black boxes.
3. Rebuild media buying around signals, not just conversions
Adweek’s “stop chasing data and start harnessing audience signals” is the right instinct, but it stops short of the operational question: what do you actually optimize on when the click and the pixel are unreliable?
Move from “click-based” to “signal-based” buying
Three layers of signals to plan and buy against:
-
Platform-native intent signals
- On social: video completions, saves, shares, profile visits, DMs started, product detail views, add-to-carts, and in-app checkouts.
- On search: impression share in key answer boxes, “Read more” link inclusion, and query mix shifts (brand + category, not just brand).
- On messaging: conversation starts, qualified lead flags, and resolution-to-purchase ratios.
Stop treating anything that isn’t a last-click purchase as “upper funnel fluff.” These are the new leading indicators.
-
Cross-channel behavioral signals
- Brand search lift after heavy in-platform education campaigns.
- Direct traffic and “dark social” referrals after viral content or PR hits.
- Retail or marketplace sales spikes after answer-engine exposure or content pushes.
This is where MMM, incrementality testing, and lightweight geo experiments earn their keep. You will not get clean path-level attribution. You can get clean pattern-level evidence.
-
Downstream quality signals
- Lead-to-opportunity conversion by source, not just lead volume.
- First-order vs. LTV by acquisition surface (social commerce vs. site vs. marketplace).
- Churn and expansion rates tied back to initial journey type where possible.
If you keep optimizing to the cheapest observable conversion, you’ll train your system to chase low-intent, low-LTV users that just happen to click.
The practical move: redefine “performance” for your media team. Give them permission and budget to optimize to high-intent signals and incrementality, not just last-click ROAS screenshots.
4. Redesign your funnel as a network, not a line
The classic linear funnel – ad → click → site → convert – is now a graph:
- AI overview → brand mentioned → user screenshots or saves → later branded search → marketplace purchase.
- Social video → comments → DM conversation → WhatsApp follow-up → offline purchase.
- Creator mention → TikTok shop add-to-cart → abandoned → remarketing in feed → in-app checkout.
None of these journeys care about your website the way your dashboards do.
Practical funnel redesign moves
- Shift from “site-centric” to “surface-centric” planning. Plan journeys per surface: search, answer engines, social feeds, messaging, marketplaces, retail. Ask “what does a complete journey look like if the user never visits our site?”
- Build conversion paths inside platforms. Use native shops, lead forms, in-app checkout, and messaging flows where they make sense. Yes, you lose some data. You gain completion rates.
- Use your site as the “deep dive,” not the only destination. Treat your domain as the place for complex configuration, account creation, and rich education – not the mandatory step for every prospect.
- Align incentives with the new reality. If your performance team is bonused solely on site-reported ROAS, they will resist anything that moves conversion into platforms, even if it’s better for the business.
The teams that win will be the ones whose org charts and incentives match how people actually buy in 2026, not how analytics reports looked in 2016.
5. Build a content and media stack that is AI-native, not AI-dependent
There’s a subtle but important difference between “we use AI tools” and “our strategy assumes AI gatekeepers.”
What an AI-native stack looks like
- Human strategy, AI-assisted production. Use AI to draft, summarize, and repurpose, but keep humans in charge of positioning, narrative, and proof. The Copyhackers warning about AI’s trust problem is real: generic AI sludge trains both humans and machines to ignore you.
- Systematic research on how you appear in AI answers. Regularly query major models and answer surfaces for your priority topics and brand. Track:
- Are you cited?
- Which competitors are?
- What sources are being referenced?
Then adjust your content, PR, and partnerships to match the patterns you see.
- Creative built for machine remixing. Short, modular, well-labeled assets (clips, quotes, images, data points) that can be reassembled into different formats by both your own tools and platform algorithms.
- Governance around AI content. Clear rules on where AI can draft vs. where humans must originate. Guardrails for claims, tone, and brand safety. You’re not just training your audience; you’re training models on what “you” sound like.
The point is not to worship AI. It’s to accept that AI systems are now distribution, discovery, and context layers and design your stack accordingly.
What to do in the next 90 days
To make this less abstract, here’s a concrete 90-day plan for a CMO or growth lead:
- Audit your “zero-click footprint.”
- Search: Where do you appear in AI overviews, People Also Ask, featured snippets, and “Read more” links?
- Social: Which posts drive saves, shares, and DMs, not just clicks?
- Messaging: How many sales-originated conversations happen in DMs or chat today?
- Define 3-5 priority topics where you must be the default answer.
- Map current rankings, AI citations, and competitor presence.
- Consolidate cannibalized content and build one authoritative hub per topic.
- Rewire one major campaign to be surface-first.
- Design separate journeys for search, social, and messaging where the user can complete without visiting your site.
- Set KPIs around high-intent signals and downstream sales, not just clicks.
- Run one incrementality test that ignores last-click.
- Geo split, holdout, or time-based test on a key channel that is currently under-credited (e.g., social video or content distribution).
- Use the results to reset how you value that channel in planning.
- Set an internal rule: “No asset ships without an answer.”
- Every ad, post, or article must clearly answer a specific question a buyer actually has.
- If it can’t be the best answer on at least one surface, it doesn’t go live.
The click is not dead. It’s just no longer the main character. The sooner your media, content, and measurement stop treating it like one, the more room you’ll have to win in the answer engine era while everyone else fights over shrinking blue links.