Should you rewrite or restructure your old content for GEO?

If you’re sitting on hundreds of old blog posts, you don’t have to rewrite them all, but you can’t just leave them as is either. The solution lies somewhere in the middle. Older content needs to be sorted into the appropriate content bucket before deciding how to use it. Treating this as one big decision to rewrite everything or touch nothing is how most Generative engine optimisation (GEO)  efforts stall before they start.

Here’s a practical way to work through it.

Why this decision matters at scale

Rewriting an entire content library isn’t realistic for most teams and it isn’t necessary either. A lot of older content is already doing a perfectly good job, it just needs a lighter touch than a full rewrite implies.

At the same time, leaving everything unchanged wastes whatever opportunity already exists in that back catalogue. Content that ranks well in Google doesn’t automatically get cited in AI search. Ahrefs has found that on average, only around 11 per cent of AI-cited URLs overlap with Google and Bing’s top ten results for the same query. Good rankings and AI citations are increasingly two separate outcomes, not one following naturally from the other.

The real risk isn’t picking the wrong overall strategy. It’s treating this as one decision instead of hundreds of smaller ones made page by page.

Step 1: Audit before you touch anything

Before deciding what to do with a page, you need three pieces of information about it:

1. Current AI citation status: Search the core question that your page answers inside ChatGPT, Perplexity and Google’s AI Overviews and note whether your page shows up as a source.
2. Existing traffic and organic authority: This is still a relevant signal for GEO. A page with strong existing authority is a better restructuring candidate than one nobody’s ever found useful.
3. Topical relevance: Is this still something your audience is actually asking about or has the underlying question moved on?

This audit stage is the part teams skip most often, jumping straight to rewriting instead. Skipping it means spending effort on pages that didn’t need it while leaving the ones that would actually move the needle untouched.

This is also, in practice, the same groundwork behind Axonn’s own GEO audits, assessing where a site currently stands across both traditional rankings and AI-driven results before recommending any changes to the content itself.

Step 2: Sort into three buckets

Once you’ve got that information, most pages fall cleanly into one of three categories:

Leave as is

This page has low traffic, low strategic relevance and no realistic path to citation, regardless of what you do to it. Not every page needs to compete for AI visibility and spending time here is effort better spent elsewhere.

Restructure

The information is accurate and still relevant, but the format is working against it. Common signs include:

• Long, unstructured paragraphs
• No clear heading hierarchy
• No direct answer near the top

This is the biggest bucket for most established blogs. The good news is that a restructure is far quicker than a rewrite, especially if you know how to structure content so AI engines can quote it.

Rewrite

This bucket is reserved for pages that fall into one of three categories:

• Genuinely outdated
• Thin on real substance
• Covering a topic your audience clearly still cares about, but doing it poorly

A full rewrite is the most expensive option, so it’s worth reserving for topics with real, ongoing value rather than defaulting to it out of caution.

A simple decision checklist

Run each page through these questions in order:

• Does this topic still matter to your audience? If no, leave it as is.
• Is the underlying information still accurate? If no, it needs at least a rewrite of the outdated sections.
• Is the content already well structured, answer-first, with clear headings and short paragraphs? If yes, it’s a strong restructure candidate rather than a rewrite.
• Is this a high-value topic not performing well on either traffic or citations? If yes, it’s a genuine rewrite candidate.

Most pages will resolve after the first two or three questions. Only a small number should end up needing a full rewrite.

These same fundamentals – content authority, structured content performance and inclusion in AI-generated summaries –  are what we track in our own GEO reporting, so the checklist above lines up with how progress actually gets measured once the work is done.

Prioritising the backlog

Once everything’s sorted, work through the backlog in this order rather than alphabetically or by publish date:

1. Quick, high-impact restructures on pages that already have decent traffic or authority
2. Full rewrites, but only for genuinely high-value topics
3. Restructures on lower-traffic pages, once the higher-impact work is done
4. Everything in the leave-as-is bucket stays exactly where it is

This order gets visible wins early, which matters if you’re trying to demonstrate GEO progress to a wider team or client, rather than disappearing into a six-month rewrite project with nothing to show for months.

The mistake to avoid

The most common GEO mistake isn’t picking the wrong bucket for an individual page. It’s treating the whole exercise as all-or-nothing, either committing to rewrite an entire site’s worth of content, which stalls out from sheer scale, or deciding GEO is too big a job and doing nothing at all.

Sorting your library into these three buckets turns an overwhelming problem into a manageable, prioritised list. Most of the value comes from the restructure bucket anyway, which is exactly the work most teams put off longest because it feels less dramatic than a full rewrite.

Want help auditing your content library?

Our GEO strategy audit sorts your existing content into exactly this framework and tells you where to start. Contact us and we’ll talk you through it.