AI didn’t kill the open web. But it certainly rewired the way people discover content and how publishers monetize. For an industry already navigating privacy changes, signal loss, and platform consolidation, AI-generated answers are the latest structural shock and the clearest signal yet that publishers need a new playbook.
The traffic shock: AI answers are cutting into web discovery
When Google launched AI Overviews in 2024, it altered search behavior. Many queries now receive an answer directly in the search results page, reducing click-throughs to publishers.
- Nearly 60% of Google searches end without a click in 2024 (i.e. zero-click), meaning users either end the session or re-query rather than following a link.
- Industry analyses show click-through declines ranging from modest to as high as 89% in certain verticals, depending on how ‘answerable’ the content is.
- In the U.S., only 360 clicks per 1,000 searches (36%) go to open-web sites. In the EU the number is 374 per 1,000.
- Bain’s “Goodbye Clicks, Hello AI” commentary warns that zero-click behavior is redefining discovery and pushing publishers to rethink how they attract audience beyond clicks.
These numbers underscore how much of search volume no longer translates into site traffic. But the impact is highly uneven:
- Strong, recognized brands with loyal audiences tend to weather the storm better. Their existing direct traffic, newsletters, or branded searches cushion losses.
- Niche/utility publishers that rely heavily on FAQ, definitions, or “how-to” content are most vulnerable, since those are precisely the queries that AI Overviews can answer.
- As Google tunes what triggers AI summaries (e.g. retracting results that mis-answer queries), some publishers see recovery or mitigation.
A two-track publisher strategy is emerging
Publishers are responding with a dual strategy, not in isolation but in parallel:
Track 1: license into AI
Publishers are engaging AI companies (OpenAI, Google, Anthropic) to license content for training or AI result consumption.
Notable public deals:
- Associated Press → OpenAI
- Reddit → Google (data licensing)
- Axel Springer → OpenAI
- Financial Times → OpenAI
- News Corp → OpenAI
Licensing models vary, including flat fees, per-query/usage, revenue-sharing, or hybrid models. But attribution (link-backs, branding) is increasingly non-negotiable for visibility and brand equity.
Source: SubStack
Track 2: own the experience
Investment in apps, CTV, newsletters, podcasts, and membership models formats where publishers control UX, signals, and user relationships. These channels are lower-leakage, meaning the user doesn’t pass through a search intermediary before consumption.
In-app and CTV are the hedge
Signal density & habit loops
- In-app environments don’t depend on search discovery. Users arrive via app stores, push, and habit loops.
- Frequent sessions, login states, device identifiers, and session metadata all combine to form rich bid signals.
- These environments also foster higher retention and repeat usage, reducing dependence on one-off traffic.
SDK enrichment & formats
- Modern SDKs (Software Developer Kits) enrich the bidstream with context (content metadata, genre), device, OS version, session data, and more.
- OM SDK adoption unlocks premium demand by proving viewability and verifying ad behavior.
- Format diversity (video, rewarded, interactive, CTV) allows publishers to monetize through engagement, not just impressions.
Tactical Playbook for Publishers
Here are concrete actions publishers should take now:
- Audit your SEO portfolio. Identify pages that are “answerable” (FAQ, definitions). These are high-risk. Reallocate into analysis, stories, and interactive content that AI can’t replicate easily.
- Enrich your SDK + adopt OM SDK. Add metadata and verification to your bidstreams. This improves bid response rates and raises CPM floors.
- Build curated PMPs. Create private marketplaces with thresholds for dwell time, scroll depth, and viewability. Sell “attention-first” inventory.
- Select signal-first SSPs/partners. Work with SSPs that support pre-filtering, predictive curation, and cross-format environments (web, app, CTV).
- Plan licensing carefully. Negotiations should address attribution, usage limitations, renewal terms, and distribution rights.
Conclusion: a reset, not a retreat
So AI hasn’t killed the open web, but it sure has rewritten the economics of discovery. For publishers, the lesson is clear: SEO-driven traffic alone can’t be the foundation of a sustainable monetization strategy.
The way forward lies in 2 complementary tracks:
- Monetize AI directly through licensing and attribution agreements.
- Invest in owned, signal-rich environments like apps, CTV, newsletters, and podcasts, where discovery is habit-driven and engagement is measurable.
The tactical playbook is emerging: enrich SDKs, adopt OM SDK for verification, package attention into PMPs, and partner with SSPs that can pre-filter noise and deliver curated supply. These steps won’t just protect yield, they’ll strengthen the link between content, users, and revenue in a signal-thin world.
But this is only half the story.
In the second part of this 2-article series, we’ll turn the spotlight to the advertisers, examining why budgets are flowing into CTV and in-app, how attention metrics are reshaping measurement, and what a high-performance supply path looks like in the AI era.