For decades, the web ran on an unspoken pact. If we provided reasonable value on our web pages, search engines would reward us with traffic, some of which turned into ad revenue, or into customers for our products and services.
That model has now been usurped by the influx of AI-extracted answers across platforms and a search experience that no longer ends at a web page.
In July 2025, Cloudflare CEO Matthew Prince shared approximate crawl ratios per referred visitor for leading platforms.
Google 18:1, OpenAI 1,500:1, and Anthropic 60,000:1.
That meant that for every visitor Anthropic sent back to a site, it crawled roughly 60,000 pages. The old exchange of content for traffic was collapsing.
What was and is being commoditized is generic, synthesis-friendly information: definitions, basic how-tos, listicles without experience, and articles that mostly restate what is already widely available. Google calls it “commodity content.”
What does this mean for the future of web content?
First, content has value when it contains original inputs that AI cannot invent: first-hand experience, proprietary research and data, expert interpretation, strong editorial judgment, and a clear point of view.
That is exactly the kind of material AI systems need in order to produce useful answers, and it is also the kind of material users still seek when they want to verify, go deeper, or make consequential decisions.
“There is one new rule in marketing. Your brand is either mentioned when buyers search, or you are invisible. They get answers before they ever click a link. If you are not in those answers, you are not in the conversation.”
— Tom Niezgoda, CMO at Positive Surfer
Second, content has value as source material and evidence, not just as a destination page. In an AI search environment, a page may create value even when the user does not read it line by line, because it becomes the page that gets cited, grounded, quoted, or used to complete a task.
Google Analytics and Microsoft Bing Webmaster Tools now both offer AI performance reporting, a strong signal that "being a source" and receiving "AI referral clicks" have become an important measurable form of search visibility.
We released Surfer AI Tracker so you can see when and where your content is being cited across AI answer engines, turning "being a source" into something you can actually measure and improve.
Third, content still has value when it is tied to trust and action. If a user is about to buy, invest, hire, choose software, follow medical advice, or learn a skill, they often want more than a summary.
They want examples that mirror their situation, comparisons, and the perspective of someone who has actually done the thing. AI can compress information, but it cannot manufacture lived experience.
The web's next phase
The web began as a read-only library of pages, slowly becoming a participatory web of blogs, forums, and social media. ChatGPT's release in November 2022 ushered in the AI-assisted web in which search, writing, and discovery were increasingly shaped by models.
We're now entering the agentic era.
AI is now expected to perform action, not just offer information. Browser-native tasks have become increasingly possible to perform directly on websites and desktop interfaces, with agents clicking, typing, navigating, and completing tasks on a user's behalf.
While Anthropic's Cowork and other early entrants opened the category, ChatGPT's worker model is rumored to be released soon. Visa and Mastercard have built agentic commerce protocols, and Robinhood even announced an agentic credit card (!).
For three decades, the web ran on the attention economy: ads, clicks, and impressions, where attention flowed to whoever could capture eyeballs. But in the agentic web, a user states a goal, delegates it, and an agent executes.
The question becomes, "Does the agent trust me enough to act through me?"
"We've never optimized for clicks for their own sake. We've optimized for being the source that matters. In a world where agents do the searching, the buying, and the deciding, that trust is everything and it's exactly what Surfer is built for.
The brands that win the next decade won't be the ones that capture the most attention. They'll be the ones agents choose to trust."
— Lucjan Suski, CEO at Positive Surfer
This breaks the classic marketing funnel. Zero-click may be nearing zero-human: the research, the comparison, and the checkout can all happen without a person ever landing on your page.
We've debated "user experience" for years. Now there's a second audience to think about: Agents. From UX to AX.
AXO? Agent experience optimization? We have enough acronyms we’re dealing with at the moment, so we don't really need another.
Open interoperability standards are becoming core infrastructure. Protocols like MCP and A2A show that the next phase of the web will depend less on one-off integrations and more on shared standards that help agents connect to tools, data, and even other agents across ecosystems.
Your website's pages can now serve as a knowledge asset that wins in two places at once: classic search results and agent-friendly visibility.
What's coming in 2026
We’ve been preparing for this at Surfer for a while.
2018 Pioneered data-backed content optimization, leading to the birth of Surfer's Content Score, now the industry standard among most content tools.
2021 Established an internal AI research team using AI and machine learning to understand how LLM models work and integrated those findings within Surfer's suite of tools.
2022 Shipped Surfer AI, the first of its kind SEO-friendly AI generative writer, before AI writing tools flooded the market.
2024 Released 1-click SEO optimization feature that included semantic internal linking, topical depth, and coverage.
2025 Merged AI search optimization with traditional SEO into a single workflow to help content show up in SERPs and in AI answers from ChatGPT and other answer engines.
We're now re-engineering Surfer's capabilities for the agentic web era.
"When AI changed everything, we had a choice: slap AI on our existing app and call it a day, or go back to the beginning and rebuild. We chose to rebuild.
In Q3, we're opening Surfer to everyone building with AI, bringing agentic experience and MCP to search visibility. No fragmented tools. No guessing. No wasted effort."
— Tom Niezgoda, CMO at Positive Surfer
Agent experience, how easily a machine can read, trust, and act on your site, will soon matter as much as how a human feels using it. Agent visibility is buyer visibility, because the agent is increasingly the one making, or at least recommending, the decision.
In practice, agent experience comes down to four things:
- Readable: content structured cleanly enough for a machine to parse.
- Citable: clear claims, sources, and provenance an agent can quote and stand behind.
- Trustworthy: verifiable expertise and first-hand experience that earn an agent's confidence.
We resolved these first 3 points with Surfer AI Tracker. But there’s another point.
- Actionable: the tools, data, and endpoints, increasingly via standards like MCP, that let an agent complete a task through you, not just read about it.
We pioneered content optimization. Now we're building for agents.
"We've spent years helping brands get found. The next decade is about getting chosen by people and by the agents acting for them. That's what we're building toward.”
— Lucjan Suski, CEO at Positive Surfer
Surfer is evolving beyond recommendations by integrating agentic workflows and MCP standards, allowing it to autonomously complete research, monitoring and optimization tasks for you.
As the internet enters a new phase, so are we, rebuilding Surfer into Surfer 3.0, a unified visibility optimization platform for the agentic web.
Start by seeing if your brand is ready for the agentic era with Surfer AI Tracker.
