Why Publishers Need an AI Strategy Before It's Too Late

The numbers are starting to come in, and they are not reassuring.

Publishers who have been watching generative AI as a distant technological curiosity are now seeing it in their analytics. Referral traffic from Google is shifting. Time-on-site metrics are moving. And in some verticals — particularly those where AI can produce a serviceable summary answer — the decline is accelerating.

The advertising implications are direct. Fewer page views mean fewer ad impressions. Fewer ad impressions mean lower absolute revenue, regardless of CPM performance. For publishers whose business model depends on scale, this is a structural problem, not a temporary blip.

What most publishers are getting wrong

The instinct is to respond to AI as a content threat — to focus on what happens to SEO, how to protect content from scraping, whether to pursue licensing deals. These are legitimate concerns. But they miss the more immediate question: what is happening to your revenue right now, and what does the trajectory look like over the next 12 to 24 months?

Most publishers I speak with do not have a clear picture of the correlation between generative AI adoption and their own ad revenue trends. They have a vague sense that something is changing, but they have not done the analytical work to understand the magnitude, the affected segments, or the timeline.

That analytical work is where the strategy begins.

The questions that need answering

Before any strategic decisions about AI licensing, content protection, or revenue diversification, publishers need to understand their own data:

• Which content categories are most exposed to AI substitution?
• What is the correlation between AI Overview prevalence and your referral traffic in those categories?
• What is the revenue concentration in your most exposed inventory?
• What does a 20%, 30%, or 40% reduction in that inventory mean for your annual numbers?

These are answerable questions. They require bringing together your web analytics, your ad server data, and some external data on AI adoption trends. The output is not a prediction — it is a diagnostic. A clear picture of where you are and what the range of outcomes looks like.

What a strategy actually looks like

An AI strategy for a publisher is not a single decision. It is a set of decisions across several areas:

Revenue diversification. If algorithmic traffic is structurally declining, the subscription and direct revenue lines become more important, not less. This has implications for the balance between advertising and subscriber experiences, for first-party data investment, and for the commercial packaging of audiences.

Content protection. There are practical steps publishers can take to make unauthorised scraping harder, and legal and licensing frameworks are evolving rapidly. Understanding your options here requires both a technical audit and a view on the emerging market for AI content licensing.

Audience development. Publishers with strong direct relationships — newsletter subscribers, registered users, loyalty programme members — are less exposed to AI-driven traffic decline than those dependent on search. Building these direct connections is the most durable response.

The common thread is that all of these decisions benefit from being made proactively, before the revenue impact becomes acute. The publishers who will navigate this transition most successfully are the ones who treat it as a planning exercise, not a crisis response.