Most people think of press as something that happens, gets a week of attention, and fades. A story runs, you share it, it slides down the feed, and that is the end of it.
I think that picture is out of date, and it matters more every month. A story about you does not fade from the places people and AI tools look. It sits there, in the same words, and gets read as if it were written yesterday. My view is that old coverage is the most underrated part of your reputation, because nobody reviews it. We celebrate new placements and never audit the ones already out there.
This post is the argument for treating your back catalog as an asset you maintain, plus an audit you can finish in an afternoon.
An article does not know what year it is
When someone asks ChatGPT, Claude, Perplexity or Google AI Mode about you, the tool is not looking at a timeline of your career. It is looking at what it can find, and an article from years ago looks the same to it as one from last month unless something on the page says otherwise.
That creates a quiet problem. The description of you that exists in the most places, or on the sources that seem most reliable, tends to win. If that description came from a profile written when you were running a different company, in a different role, in a different city, the tool has no reason to prefer your current story.
I can use my own history as an example, with the caveat that I am describing the shape of the risk and not claiming a specific bad result. Before Arizona I worked in tech sales and startup marketing around Seattle. That chapter is real and I am proud of it, and it is also not what I do now. If the loudest descriptions of me were all from that era, a person asking about me today would get the wrong picture, and nobody would have lied. The information would just be old.
That is the pattern I want you to watch for. Old coverage rarely says something false. It says something that used to be true.
Three kinds of old press, and what each one costs you
When I look at coverage with this in mind, I sort it into three buckets.
Stale but harmless. A profile or mention that describes you accurately enough, even if it is dated. It might say you work in a field you have since narrowed. This is fine. It still counts as independent proof that you exist and are taken seriously.
Stale and misleading. The title is wrong, the company is wrong, the service list is something you stopped offering, or the story frames you around a chapter you moved on from. This is the bucket that does damage, because a tool repeating it sounds confident and a reader has no clue it is out of date.
Old and unflattering. Something you would rather nobody read first. A complaint, a bad review that got written up, a story about a rough patch. I will be straight with you here. I do not think the answer is to panic or to try to scrub the internet. Most of these cannot be removed, and chasing removal can make the thing louder. The answer is usually to make sure newer, accurate, independent material exists alongside it so the old item is one voice among several instead of the only one.
Notice that only one of the three buckets calls for urgent action, and most people never find out which bucket their coverage is in because they never look.
Why this is getting more important, not less
There are two reasons I think the stakes are rising.
First, answers are getting shorter. A page of search results lets a person scroll, compare dates and make up their own mind. An answer from an AI tool compresses all of that into a couple of sentences. The compression strips out the dates, and with them the clue that a source is old.
Second, the sources AI tools lean on are often the ones that are easy to find and read, and a lot of old articles are exactly that. They have been sitting on stable pages for years. A fresh page has to earn its way into the answer. An old one is already there.
I wrote earlier about why being featured is not the same as being found. This is the other side of that coin. Being found by the wrong version of your story is its own kind of problem, and it is the one that is easy to miss because the coverage looks like a win on your own shelf.
I do not have a clean number for how much weight any tool puts on article age, and I would be suspicious of anyone who says they do. These systems change, and they do not behave identically. What I am confident about is the direction. When the most available description of you is old, answers about you drift old too.
The afternoon audit
Block out two or three hours. You need a spreadsheet and a private browser window.
- List everything. Search your name, your company name and each of your former company names. Add every article, interview, directory entry and profile that mentions you. Go three pages deep, not one.
- Record the facts each one states. Your title, your company, your location, your services, any numbers. One row per source.
- Mark each row as harmless, misleading or unflattering, using the three buckets above.
- Ask AI. Put the same plain question to ChatGPT, Claude, Perplexity and Google AI Mode: who is this person and what are they known for. Save the answers next to your spreadsheet.
- Trace the errors. For each wrong or dated detail in an answer, find the row that most likely taught it. You will often find that a single old page is behind most of it.
- Pick the fixes. For pages you control, update them. For pages you do not, ask the publisher or author for a correction, politely and with the current facts in hand. Many will say yes to a short, specific request, and some will not respond at all.
The fixes you cannot get are the reason for the second half of the work, which is adding fresh, accurate, independent coverage. A newer article that states your current role plainly gives every tool something better to quote. That is the real argument for steady, quality press over a one-time burst, and it ties into why one great article beats ten forgettable ones.
What I would do this week
Do step one and step four only. Just list what exists and ask AI the question. You will probably learn more from those two hours than from a month of planning, and you will know which bucket you are in.
Then answer one thing honestly. If a stranger read the oldest article about you tomorrow, would they come away with the right idea of who you are now? If the answer is no, you have found your first fix, and it is probably a lot cheaper than a new campaign.
