Products SEO & GEO Agent Paid Ads Agent Website Agent Conversion Agent
How it works Pricing FAQ Blog Book a call

Refresh content with AI: update old pages for SEO faster

Updated · 7 min read

Short answer AI speeds up three parts of a content refresh: sorting your Search Console data to pick pages, finding gaps and drafting rewrites, and suggesting titles and meta descriptions. It can't supply your prices, photos or job experience, and it invents facts. Google says to fact-check all AI output, titles included, and never refresh a page just by changing its date.

Want help applying this to your business? Talk it through on a free 30-minute call.

Book a call

What AI can do in a content refresh, and what stays with you

AI is fast at sorting, spotting gaps and drafting. It is unreliable at facts. Google explains why:

"Generative models don't retrieve facts, but predict a likely sequence of words based on their training data." (Google Search Central: Guidance on using generative AI content)
JobLet AI do itYou check or supply
Pick pages to refreshSort your Search Console export and flag patternsThat the pages still match services you sell
Find what a page is missingCompare the page with the searches it getsWhich customer questions are real
RewriteDraft clearer sections, shorter sentences, an FAQEvery price, rule, date, license and claim
Titles and descriptionsOffer five options per pageThe one that's true and specific
Internal linksSuggest which of your pages should link to whichThat each link helps a reader
ExperienceNothingReal photos, job stories, your own numbers

Deciding whether a page should be updated, merged or deleted at all is covered in content decay: update, merge or delete. This guide is the AI workflow once you've decided to update.

Step 1: Let AI pick the pages worth refreshing

Start from data, not from a hunch. In Google Search Console, open Performance, set the date to compare the last 3 months year over year (Google Search Central: Debugging traffic drops), open the Pages tab and export it. The screen export holds up to 1,000 rows (Google Search Central Blog), more than a small business site needs. New to the tool? See Search Console for a small business.

Paste the table into an AI chat and ask it to sort pages into three groups:

  • Falling: clicks down a lot against last year.
  • Seen but skipped: many impressions (times shown in search) and a low CTR (click-through rate, the share who click).
  • Almost there: average position between about 8 and 20, the bottom of page one or the top of page two.

The export holds search data, not customer records. Don't paste client records or private files into a chat tool.

Some sites have a built-in option. In December 2025 Google began rolling out AI-powered configuration in the Performance report, "to a limited set of websites" at first. You type a request such as comparing blog pages this quarter with last year, and it sets the filters. Google's own caveat:

"AI can sometimes misinterpret requests. Always review the suggested filters to ensure they match your intention before analyzing the data." (Google Search Central Blog, December 4, 2025)

It can't sort or export the table for you (same post), so the paste step above still applies.

Step 2: Ask AI what each page is missing

Give the AI three things: the page text, the top searches from that page's Queries tab in Search Console, and two or three questions customers asked you this year. Then ask:

  1. Which of these searches does the page not answer?
  2. Which sentences sound dated (old years, old prices, products or rules that may have changed)?
  3. What would a first-time customer still need to know before calling?

Also search the page's main query yourself. If the top results answer a different need (prices where you explain, a checklist where you tell a story), tell the AI so it can reshape the page.

Treat its answers as a to-do list, not as facts. If it says a rule changed, look the rule up at the source before you write a word.

Step 3: Rewrite with AI, then run these checks

Have the AI draft the new sections in your words: paste one or two pages you wrote yourself and ask it to match that tone. Then go line by line before publishing. Google is direct about this:

"It is critical to manually factcheck and review all AI-generated content for accuracy and trustworthiness before publishing." (Google Search Central: Guidance on using generative AI content)

Before you hit publish, confirm:

  • Every number is yours or sourced. Prices, warranty terms, response times, permit rules. AI fills gaps with plausible guesses.
  • Something only you could add is there. A photo from a real job, what usually goes wrong, the question every customer asks. AI can't invent that honestly.
  • The page answers one customer, start to finish. Cut padding the AI added.
  • Links work and point where the text says.
  • A visitor can tell how it was made, if they'd wonder. Google suggests disclosing AI use where someone might ask "How was this created?" (Google Search Central: helpful content).

Refresh a few pages well rather than all of them at once. Google's spam policy lists "using generative AI tools or other similar tools to generate many pages without adding value for users" as scaled content abuse (Google Search Central: Spam policies). More on the policy: does Google penalize AI content?.

Ask for five title options and five descriptions per page, then pick and edit. Google's review rule covers these too: the fact-check "also applies to metadata like title elements, meta description elements, structured data, and alternate texts for images" (Google).

  • Title: Google asks for "descriptive and concise text" and warns against long, verbose titles (Google Search Central: title links). Name the service and the city if you serve one.
  • Meta description: make it unique; "identical or similar descriptions on every page of a site aren't helpful" (Google Search Central: snippets). Google writes snippets mainly from the page itself, so a good description helps but isn't guaranteed to show (same page).
  • Internal links: ask the AI which of your other pages mention the topic, then add links with clear link text. Google says link text "tells users and Google something about the page you're linking to" (SEO Starter Guide).

For what a service page should contain, see on-page SEO for service pages.

Step 5: Republish, ask Google to recrawl, then measure

Show a "Last updated" date only when the content truly changed. Google lists this as a warning sign:

"Are you changing the date of pages to make them seem fresh when the content has not substantially changed?" (Google Search Central: helpful content)

Then request indexing in Search Console's URL Inspection tool for a few pages, or resubmit your sitemap for many. Google says recrawling "can take anywhere from a few days to a few weeks", and asking again for the same address "won't get it crawled any faster" (Google Search Central: recrawl).

Wait about a month, then compare the refreshed pages' clicks, impressions and position with the same weeks before. Track calls and form fills too, not just traffic: is my SEO working?.

Example

Illustrative cases with invented numbers.

A roofer with 30 blog posts (US dollars)

The export shows 4 posts with 2,000+ impressions a year and under 1% CTR. AI drafts new titles and an FAQ for "roof replacement cost". The owner swaps the AI's guessed $8,000–$12,000 range for his real 2026 range and adds photos from two jobs. He requests indexing for the four pages only.

A physiotherapy clinic (Canadian dollars)

AI flags a page saying a session costs C$80, from 2022. The clinic updates it to its current C$110 fee, and a physiotherapist checks the AI's rewrite of the treatment section line by line before it goes live.

A cleaning company that refreshed 200 pages in a day

It had AI rewrite every city page and set all dates to today. No new facts or photos were added. That is the pattern Google's spam policy and date warning describe; a few real updates would have served it better.

FAQ

Is it safe to use ChatGPT to rewrite old blog posts?

Yes, if a person checks every fact and adds what only the business knows. Google's concern is content made mainly to rank or mass-produced without value, not the tool.

Will updating the date on a post help it rank?

Not on its own. Google calls changing dates without substantial changes a warning sign of search-first content.

How often should I refresh content?

When the facts change or Search Console shows a page falling. For most small sites, a quarterly look at the year-over-year Pages report is enough.

Can AI write my meta descriptions?

It can draft them. Google says to review AI-written titles and descriptions like any other content, and to make each page's description unique.

How long until a refreshed page moves?

Google says recrawling takes days to weeks, and ranking changes, if any, come after that.

Different for you?

Keep learning

NextSite content analysis: how to review every page on your siteA site content analysis in one spreadsheet: list every page, add Google clicks and leads, score each one, and decide to keep, update, merge or delete it.

Ask us one question

One line is enough. We reply by email.

Sources

  1. Google Search Central: Guidance on using generative AI content on your website
  2. Google Search Central: Creating helpful, reliable, people-first content
  3. Google Search Central: Spam policies for Google web search
  4. Google Search Central Blog: AI-powered configuration in Search Console (December 4, 2025)
  5. Google Search Central Blog: Performance data filtering and limits (October 2022)
  6. Google Search Central: Debugging drops in Google Search traffic
  7. Google Search Central: Influencing title links in Google Search
  8. Google Search Central: Control your snippets in search results
  9. Google Search Central: SEO Starter Guide
  10. Google Search Central: Ask Google to recrawl your URLs