The quiet shift that’s breaking your media plans
Look at those headlines again and a pattern jumps out: everyone is still obsessing over micro-optimizations (title tags, local landing pages, TikTok posting times, social schedulers, conferences) while the actual battleground is moving somewhere else.
That “somewhere else” is simple to describe and hard to operate in:
- AI systems summarizing, rewriting, and answering instead of sending traffic
- Platforms fighting for time and habit, not just clicks and views
- Users swimming in AI-generated slop, tuning out anything that feels generic
- Media and commerce giants building their own closed content and ad ecosystems
Ben Thompson calls it the attention economy. That’s not new. What is new is that the attention economy is now AI-mediated, platform-owned, and increasingly click-optional.
If you’re still running a marketing and media strategy that assumes:
- Search = “10 blue links + ads + organic clicks”
- Social = “feeds that send traffic to your site”
- Video = “views that drive site visits or branded search”
…you’re optimizing for a web that’s disappearing.
Three shifts you can’t ignore (even if your dashboards still look fine)
1. AI intermediaries are eating the “consideration” layer
Microsoft is already warning that “summarize with AI” buttons are poisoning recommendations. That’s the polite way of saying: users are handing their intent to AI layers that sit between your content and their decision.
At the same time:
- Search engines are pushing generative answers above traditional results
- Agentic AI tools are being wired into social workflows and research habits
- Vector search and transformers are turning your content into embeddings, not visits
Your content is no longer just read by humans. It’s being digested by models that:
- Summarize you into a sentence inside someone else’s UI
- Blend your message with competitors in a single answer
- Strip out your brand, formatting, and CTAs as “noise”
The old game: “rank, get the click, convert.”
The new game: “be the reference the model trusts, even if the user never clicks.”
2. Platforms are optimizing for time spent, not traffic sent
Look at the surface area:
- YouTube rolling out new business tools and AI video creation
- TikTok best-time-to-post research and scheduling tools
- Bluesky, decentralized social, and new walled gardens
- Streaming rankings (Stranger Things, etc.) treated like Nielsen for the new era
- Retail media and principal media gymnastics in programmatic
Every platform is chasing the same KPI: keep users inside, keep them scrolling, keep them shopping without leaving.
That means:
- Fewer outbound clicks per impression
- More native formats (shopping, lead forms, in-app checkout, in-stream actions)
- More “content as container” for commerce, not just awareness
If your media model still assumes “we buy impressions here to get visits there,” you’re fighting the platforms’ business model, not working with it.
3. AI slop is raising the bar for anything that looks automated
TechCrunch is asking if the creator economy can even stay afloat in a flood of AI slop. Copyhackers is talking about AI’s trust problem and broken email experiences. Depop is being praised for not drowning in AI-generated junk images.
Users don’t know what’s AI-generated, but they can feel what’s lazy:
- Samey landing pages with identical hero layouts and “value prop” jargon
- Emails that technically personalize but emotionally say nothing
- Videos that look like every other template in the tool’s library
In a world where anyone can spin up content, the scarce resource is not production – it’s signal. Distinctive voice, sharp POV, and proof that a real brain is behind the message.
What this actually means for CMOs and media leaders
You don’t need another think piece telling you “AI is big” or “attention is scarce.” You need operating rules that change where you put money, people, and time.
Rule 1: Plan for “zero-click” impact, not just click-through
Your reporting stack probably still treats “no click” as “no impact.” That’s wrong in an AI-first, platform-owned environment.
You need to measure and design for:
- View-through and search-lift as primary signals
Treat branded search volume, direct traffic, and category search share as core media KPIs, not “nice to have.” - On-platform outcomes
Lead-gen forms, add-to-cart in retail media, in-app subscriptions, in-feed shop actions – these are not side quests. They’re your new “landing pages.” - Assist value of content that never gets clicked
If your content is quoted, summarized, or frequently surfaced in AI answers or social conversations, that’s distribution – even if your sessions graph doesn’t spike.
Practically:
- Rebuild your attribution view to include impression-level and view-through data as first-class citizens
- Hold channels accountable for incremental lift in branded/search demand, not just last-click ROAS
- Stop killing “upper-funnel” work that clearly moves intent but doesn’t show up in your current MTA model
Rule 2: Design content for machines and humans
The SEO crowd is already talking about “generative engine optimization.” Strip away the buzzword and the job is:
- Give models clean, unambiguous, factual, up-to-date signals
- Give humans a reason to care once they actually land on you
What this looks like in practice:
- Structured clarity
Use schema, clear headings, concise definitions, and canonical answers. Make it obvious what your page is about and what you’re the authority on. Models love clarity. - Distinctive POV layered on top
Don’t stop at “what is X?” Add “here’s the trade-off,” “here’s the failure mode,” “here’s what everyone gets wrong.” This is what humans remember and share. - Source-worthiness
Publish data, frameworks, and original research that AI systems and journalists can cite. You want to be the table everyone copies from.
If your content can’t answer: “Why would an AI model pick this as the reference example?” you’re writing for a shrinking audience.
Rule 3: Move budget closer to where attention actually lives
Look at where the industry is investing attention:
- Retail media and Merchant Center feeds (and their disruptions)
- YouTube and streaming ratings as the new “prime time”
- Social schedulers, TikTok timing, and always-on short video
- Agentic AI inside social and CRM workflows
Then look at your own budget mix. If you’re still over-indexed on:
- Driving traffic to homepages that don’t convert
- Generic search campaigns with no creative or audience strategy
- “Set and forget” performance media that assumes stable CPC/CPA dynamics
…you’re funding a nostalgia project.
Concrete moves:
- Shift from channel-first to surface-first planning
Instead of “our Google budget” or “our Meta budget,” think “our product card surfaces,” “our short video surfaces,” “our AI-answer surfaces,” “our email surfaces.” Then fund the surfaces that actually get seen. - Invest in native conversion, not just traffic
Build proper product feeds, creative variations, and offer logic for retail media, social shops, and YouTube CTAs. Treat these like your new storefronts. - Protect experimentation budget
New ad surfaces (AI search answers, shoppable streams, decentralized social) will be underpriced and under-measured at first. Ring-fence 5-10% of spend to learn there before your competitors do.
Rule 4: Treat creative and narrative as performance infrastructure
The PPC headlines are already pointing at the future: AI, visual creative, and new ad surfaces. Meanwhile, a lot of performance teams are still behaving like creative is a service desk ticket.
In an AI-mediated attention economy, creative is not “the last step.” It’s the only thing the user ever actually sees.
For operators, that means:
- Shorten the creative feedback loop
Build small, cross-functional pods where media buyers, data analysts, and creatives sit on the same weekly review, looking at the same performance data. - Use AI for volume, humans for edge
Let AI generate variations, resize assets, and test angles. But insist that humans define the core narrative, voice, and “we’re willing to say this and not that” line. - Codify what works
Turn winning hooks, angles, and formats into playbooks. Not brand guidelines. Actual examples: “This is the 6-second pattern interrupt that cut CPV by 30%,” “This is the email structure that fixed our 73% broken state.”
Rule 5: Build trust as a measurable asset
AI’s trust problem isn’t just an AI problem. It’s your problem when:
- Prospects can’t tell if your message is written by a bot
- Influencers and digital twins are selling on your behalf
- Users are burned out on fake personalization and over-automation
Trust used to be a brand slide. Now it’s a performance variable.
Ways to operationalize it:
- Make “realness” visible
Show who wrote the thing. Show their face. Show their credentials. Show what they actually believe. This matters in content, in video, in email, and even in ad creative. - Audit your automation
If 73% of your emails are broken, your “personalization” is probably doing more harm than good. Run a ruthless audit: what would this look like if I received it as a customer? - Use community as proof, not as a vanity metric
Email communities, social listening, and reviews are not just “engagement.” They are live, public, machine-readable signals that you are who you say you are.
How to know if you’re still playing the old game
A quick diagnostic you can run in your next leadership meeting:
- Dashboard test: If your main marketing dashboard is 90% click-based and 0% attention or intent-based, you’re flying with old instruments.
- Org chart test: If creative, content, and media sit in different silos with different goals, you’re not set up for an attention-first world.
- Budget test: If less than 10-15% of your spend is on native on-platform conversion surfaces (retail media, in-app shops, lead forms, shoppable video), you’re under-exposed to where users actually act.
- AI test: If your AI usage is mostly “write this blog post for me” and not “help us analyze patterns, generate variations, and feed insights back into strategy,” you’re using a jet engine as a paperweight.
The uncomfortable, useful question
Every CMO and performance leader should be asking:
“If my traffic dropped by 30% because AI and platforms kept more of it, would my marketing still work?”
If the answer is no, you don’t have a performance problem. You have an attention economy problem.
The fix is not another round of bid optimizations or another conference about SEO trends. It’s rebuilding your strategy around where attention actually lives now: inside AI answers, inside platform-native experiences, and inside the small set of messages that feel real in a world full of slop.