Stop Letting Old Articles Define Your AI Search Presence
AI search presence come mostly from press, reviews and competitor content. Here's how to check what gets cited each month and respond while coverage is current.
A buyer at a mid-size company asks an AI assistant for vendors in your category. Good news: you make the list. You’re also described from a two-year-old article, with your old product name and a feature you retired last spring. No one has looked at that article since it ran, because your team isn’t doing the kind of competitor monitoring that would catch it.
An hour later your CMO forwards the screenshot with one line: “WHY IS THIS WHAT AI SAYS ABOUT US!?!?”
You’ve sat through the conference session on getting found in AI search, but you still don’t have a way to check what the answers say or where they come from.
Why the AI Answer Is the First Impression
People read the summary and move on. The data backs it up. When Google showed an AI summary, U.S. adults clicked a search result in about half as many visits, according to Pew Research Center’s tracking of their searches.
Pew measured everyday search behavior. Still, picture someone building a shortlist. The paragraph about you in that answer may be the only description of your company they read before the first call. If it’s two years stale, you can drop off the list before sales knows the deal existed.
The buyer won’t tell you, either. There’s no form fill for “read an AI answer and moved on.”
Where AI Search Answers Get Their Material
The answer your buyer read was stitched together from trade press, analyst notes, review sites, comparison posts, your competitors’ content and, somewhere in there, your own site.
Each source plays a part. A review site supplies the pros and cons. A trade article supplies the product description, fixed on the day it ran. An analyst note supplies the category you’re filed under. The answer blends them, and the reader can’t see how old any of them are without clicking through, which most people don’t.
Your site is a small slice of it. When University of Toronto researchers compared sources for U.S. software queries, AI search drew 72.7% of its citations from independent media, review and comparison sites, and 26.7% from brand sites. The study also covered consumer electronics and cars, and security vendors sit closest to the software results.
Your competitors’ pages are in that mix. Their “us vs. them” page and the coverage of their last launch are material for answers about you.
Your own pages still count when they say something new. Original content is what gets cited, so a page that repeats the category’s talking points adds little.
Even a great homepage rewrite only reaches about a quarter of the sources. Everything else AI search draws on was published by someone else.
Remember that when someone pitches a Generative Engine Optimization (GEO) project that starts and ends on your own site. Ask what it does about the review that still lists your retired feature.
Changing What the Coverage Says
Start by knowing what’s published about you and your competitors the week it goes up, while you can still do something about it.
When a fact is out of date, fix it where it at the source. Ask the publication to update the article. Send the review site your current product details. Brief the analyst who covers you. It can feel a little passive-aggressive to email someone about their own page, but most editors and analysts don’t want stale facts on their site either. They’ll be glad you reached out (and flattered you read their article).
Keep the request short and specific: “Your review from two years ago lists our old product name. Thanks for including and covering the category. Here’s the current name and feature list, if you wouldn’t mind updating.” Editors can act on a note like that in a few minutes.
When a competitor publishes a comparison post, answer it with your own angle. Make it specific enough that a writer covering the category would quote you.
It’s easy to start with a rebuttal mindset, but skip it. Name the buyer’s real question, show where each option fits, and be honest about where the competitor is stronger. Writers link to the post that helps their reader decide.
And publish things other people can point to: original data, or a clear position on a question your buyers argue about. The effort shows up in the angle and the specifics, which is what makes a post worth citing.
Nobody controls when a model picks up a new article, so the answers catch up on their own schedule. The next round of coverage, and what AI search builds from it, describes the company you are now.
A Monthly AI Search Check Your Team Can Run
Seeing what the answers say takes the same questions every month and somewhere to write down what comes back.
- Write five to ten questions a buyer would ask: the category question, “[your company] vs. [competitor],” and “alternatives to [competitor].”
- Run them in the AI assistants your buyers use, with the same wording every month. Rewording changes which sources come back. Start a fresh session each time so earlier chats don’t shape the answer.
- Log every cited source in a shared sheet: who published it, when, and which company it’s about.
- Mark the ones your team could act on, whether that’s a correction, a briefing or a response post.
After a few months, the log starts to show patterns. Maybe one comparison post is cited in every answer about you and a particular competitor. That’s your first response post. Maybe a review site keeps turning up with old pricing. That’s an email to its editor. Maybe your own blog never appears at all, which means it isn’t saying anything worth quoting yet.
Compare each month’s log with the last. A new source is coverage someone published since your last check. A source that dropped out may have been updated or replaced. Either way, the log tracks the material, which is the part your team can change.
Keep it to about an hour. Treat AI search as one signal among several, next to pipeline, traffic and what sales hears on calls.
Tracking the Coverage Between Checks
Chatter fills the weeks in between, when new coverage is still current enough to answer. Add up to three competitors to your brand profile, and Stories shows what they publish and what’s written about them, filtered to your keywords and personas. When you prep the monthly check, filter Stories to competitor content only, so you know which new articles might turn up in the answers.
Say a competitor launches a new module on a Tuesday and a trade publication covers it on Wednesday. If that coverage reaches your team in Stories that week, you can publish your take while writers are still on the story. Found a month later in your check, it’s already part of the answers.
Every story links to the original, so your team reads the coverage itself before deciding how to respond. A story worth answering becomes a Topic, a brief you can export and draft from. Chatter isn’t in the business of writing your content for you.
It also doesn’t query AI assistants or measure AI visibility. The check stays with your team. Competitor tracking is part of the Pro and Enterprise plans.
Back to the Screenshot
This time you can answer your CMO. You know which article the answer came from, you’ve asked the publication to update it, and your response to the competitor’s comparison post is in the brief. Next month you rerun the same questions and see what’s cited now. If that two-year-old article is still there, it’s time for a follow-up.
AI search is coverage your team can watch and answer. Start this week: run five buyer questions in an AI assistant and write down every source it cites.