Refreshing what you have, usually beats publishing more.

The instinct when visibility is poor is to publish. Frequently the better move is to take pages that already have authority and make them usable, because a page that already ranks and cannot be quoted is a solved problem waiting for an afternoon of work.

Why refreshing wins

New content starts from zero on authority, links and history. An existing page that ranks has already earned those. If it is not being cited, the constraint is usually format rather than merit.

Refreshing also compounds: the same page can be improved repeatedly as you learn what gets quoted, whereas each new post starts the cycle again.

Choosing what to refresh

• Pages that rank but are never cited. The clearest signal that structure, not authority, is the problem.
• Pages answering questions your buyers actually ask. Check against your prompt set.
• Pages with stale specifics. Outdated prices, superseded processes, retired services. These are worse than absent, because they can be quoted wrongly.
• Your highest commercial-value pages, regardless of current performance.

A refresh checklist for one page

Work through the page in this order, which puts the highest-value changes first:

• Check every fact: prices, dates, service areas, staff names, product names, statistics and their sources.
• Rewrite the first sentence under each heading so it states the answer rather than introducing it.
• Rename vague headings as the questions the section answers.
• Add specifics where there are adjectives: ranges, timeframes, conditions.
• Update or add structured data, and make sure it matches the visible page.
• Add a visible “last updated” date and change the dateModified in the markup, only if the update was substantive.

What a refresh involves

Restructure for extraction: question-shaped headings, answers stated first, sections that stand alone. Replace hedged claims with specifics. Add or correct schema. Update anything time-sensitive and say when it was updated.

Then leave it alone long enough to judge. Refreshing the same page repeatedly within weeks produces noise rather than signal, and you will not know which change did what.

The stale-information risk

One argument for refreshing that is specific to AI search: a model that has absorbed an outdated claim about you will repeat it confidently. An old price, a service you dropped, an address you left.

Correcting the source is the only fix, and it takes time to propagate. That makes auditing your existing content for stale specifics more urgent than it used to be, because the cost of being wrong is now someone being told the wrong thing rather than merely reading it.

Getting the refreshed page recrawled

A refresh does nothing until the engines see it. Request indexing in Google Search Console for the updated URL, and submit it through IndexNow if your platform supports it; Bing, and therefore Copilot, uses IndexNow to pick up changes quickly. Make sure the page is in your XML sitemap with an accurate lastmod date.

Then note the date in your monitoring sheet. When you next run your prompt set, you will know which answers could have been affected by the change, and which could not.

Refresh or rewrite?

Sometimes a page is beyond refreshing. If the topic has changed enough that most of the page is wrong, if the page targets a question nobody asks anymore, or if two pages overlap so much that they compete, a rewrite or consolidation is the better move.

When consolidating, keep the URL with the most links and history, merge the useful content into it, and 301-redirect the other. Do not delete pages with links or traffic without a redirect; that throws away the authority that made refreshing attractive in the first place.

Frequently asked questions

Is refreshing content better than writing new content for AI search?

Usually, yes. Existing pages already carry authority and history; if they rank but are not cited, the constraint is format rather than merit, and restructuring is faster and cheaper than starting from zero.

Which pages should I refresh first?

Pages that rank but are never cited, pages answering questions in your prompt set, pages with stale specifics such as outdated prices, and your highest commercial-value pages regardless of current performance.

How often should I refresh content?

When there is a reason: stale information, a structural problem, or a measured gap. Refreshing the same page repeatedly within weeks makes it impossible to tell which change produced which result.

Should I change the published date when I refresh a page?

Keep the original published date and update the modified date, visibly and in structured data, when the change is substantive. Changing dates without real changes is misleading and easily noticed.

How do I know if a refresh worked?

Compare the page's citations on your fixed prompt set, its Search Console clicks and impressions, and its AI referral traffic for the weeks before and after the change.