30-second rundown
Key learnings:
- Stop judging content on traffic alone. In AI Search, a page can be valuable because it earns citations, strengthens brand associations, supports a conversion, or does all three.
- Prioritize what AI cannot finish for the user. Original evidence, live information, meaningful comparisons, official facts and task-completion pages retain value after the summary has already been served.
- Generic information is becoming a bad investment. Definitions, commodity guides and templated pages built from public facts are easier for AI interfaces to absorb and replace.
This week: Take the next ten content ideas in your queue and score each one for click resilience, citation potential, business value and proprietary advantage. Kill or rewrite anything whose main value can already be summarized from public information.
Coming 3 months: Build this scoring model into your editorial workflow and automate a monthly review using your AI visibility, analytics and conversion data. Let the system flag content types that keep earning citations, qualified visits or conversions and steadily reduce investment in the formats that do none of the three.
Aleyda Solis has walked into the content factory with a six-point inspection sheet and a can of gasoline.
The old machine rewarded volume, rankings and clicks. The new one is stranger. AI can answer the easy question before the user ever reaches your page, which means the question is no longer simply whether content can rank. The question is whether the page still has a job after the machine has spoken.
1. Traffic, Citations and Business Value Are Three Different Animals
Solis starts with a crucial distinction: AI traffic and AI citations are not the same thing. Her broader analysis of 40 leading sites across four verticals found that AI referrals are only a fraction of organic search traffic, while pages receiving AI visits and pages cited in AI answers often serve different roles.
A proprietary study may become citation bait without sending much traffic. A pricing, booking or account page may generate valuable visits without ever becoming a citation source. If those outcomes are thrown into one bucket, the spreadsheet becomes a hallucination with gridlines.
Her framework therefore separates six dimensions: click resilience, citation potential, brand mention potential, business value, proprietary advantage and expected effort. Content has to be judged by the role it can realistically play, not by one universal KPI.
2. The Best Content Leaves Something Unresolved
AI is strongest when the user wants a quick explanation assembled from common knowledge. It is weaker when the next step requires proof, live data, context, judgment, personalization or an action that must happen somewhere else.
That is why Solis prioritizes pages that remain useful beyond the answer itself: official brand and entity pages, transaction pages, product documentation, first-hand tests, original research and live first-party databases. These pages contain facts, evidence or functionality that cannot be cheaply reconstructed from ten other websites.
Click resilience rises when the user still needs the site to verify something, inspect evidence or complete an action. Citation potential rises when the page contains unique, verifiable or primary information. Proprietary advantage rises when competitors cannot reproduce the thing without your data, access, systems or expertise.
3. Generic Content Is Becoming Industrial Waste
The other side of the framework is where the bodies are buried. Content that merely restates widely available information has a weak reason to exist once an AI interface can summarize the same material instantly.
Solis does not argue that every low-priority page should be deleted. Some should be improved with original value, some consolidated into stronger resources and some removed only after considering their current performance, user role and technical implications.
But the principle is savage and useful: another generic definition, commodity guide or templated page is not automatically an asset just because there is search volume attached to it. If AI can fully satisfy the need from information already sloshing around the open web, the page needs another reason to live.
The editorial budget now belongs where there is unfinished business.
Build the thing the machine still needs to cite, the user still needs to visit, or the business still needs to convert. Everything else deserves interrogation before another dollar is poured into it. In the AI Search era, publishing more is not strategy. Publishing what still has a job is.
