SEO Is Changing, and So Is the Way I Write Content

For years, writing for SEO came with a checklist. Pick a keyword, repeat it enough times, hit a word count, sprinkle in some headings, publish. If you've ever hired a writer (or been one), you've probably seen that checklist in action.

Search doesn't work like that anymore. People now get answers straight from Google's AI Overviews, ChatGPT and Perplexity, sometimes without clicking anything at all. So I've changed how I write content, and in this post I'll show you exactly what changed and why, so you can do the same on your own site.

What changed in search (and what didn't)

Two things changed at the same time: where people find answers, and how often they click.

In March 2025, Pew Research Center tracked the browsing of 900 US adults. About 18% of their Google searches showed an AI summary. When a summary appeared, people clicked a regular search result on 8% of visits, compared with 15% when there was no summary. Clicks on links inside the summary itself? Just 1% of visits.

That sounds scary if your business depends on search traffic. But this next part is what calmed me down.

Google's guide to optimizing for generative AI features says SEO best practices still apply, because AI Overviews and AI Mode are built on the same core ranking and quality systems as regular search. The AI pulls pages from Google's index, reads them, and links the ones that support its answer. Good pages still win. The reward just shows up in more places now.

Go deeper: grounding and query fan-out

Google's guide names two techniques behind its AI answers. The first is retrieval-augmented generation, which Google also calls grounding: its ranking systems fetch relevant, current pages from the index, and the AI writes its response from those pages, with links back to them.

The second is query fan-out. The model runs several related searches at once to fill in the answer. Google's example: a question about a weedy lawn might also trigger searches about herbicides and removing weeds without chemicals. In practice, your page can be pulled in for a sub-question even when it doesn't rank for the main query. Google also warns against making a separate page for every possible fan-out query, so cover the topic well on one page instead.

👉 The rules didn't get replaced. The reward for following them well now shows up in AI answers too.

Change #1: I answer the question first

The first sentence under every heading I write now answers that heading. The explanation comes after it, not before.

A lot of SEO content (some of mine included) used to open with a warm-up: background, a dictionary definition, a "since the dawn of the internet" intro. The actual answer showed up in paragraph four. Readers don't wait that long anymore, and honestly, they never really did.

Here's how my writing habits compare now:

Why does this help? People skim, and they decide in seconds whether your page has their answer. A clear answer is also easier for Google to lift into a featured snippet, and easier for an AI system to understand and cite.

To be fair, Google says you don't need to write in a special way for AI search, and they're right. I don't do this for robots. I do it because it's kinder to the reader. Helping with snippets and AI answers is a bonus.

Change #2: I check every claim against the source

Before I write how Google or any AI platform works, I find the official page that says it, read it myself, and link it.

I started doing this because SEO is full of folklore. "Posts need 2,000 words." "Aim for 2% keyword density." "You need a special file so AI can read your site." People repeat these with total confidence, and most of them are guesses that hardened into rules.

Take word count. Google's helpful content guidance lists "writing to a particular word count" as a warning sign, and adds that Google doesn't have a preferred one. That's one link that ends a long-running argument.

Not everything in SEO is documented, though. So I sort what I write into three buckets and word each one differently:

• Confirmed. The platform's own documentation says it, and I link it.
• Seen in my work. Patterns I notice across sites I audit, clearly labeled as my experience.
• Debated. Things many SEOs believe that nobody has confirmed. I say so plainly.

This matters more than it used to. The same Google guidance asks whether content shows first-hand expertise and real depth of knowledge, and whether a reader would leave satisfied. A page full of recycled myths fails both questions.

Change #3: I write for AI answers, not just rankings

Writing for AI answers comes down to two things: your page has to be reachable, and it has to be worth quoting. That's the question I get asked most, and the answer is less mysterious than people expect.

Being reachable

An AI tool can only use your page if it can get to it. Google says a page has to be indexed and eligible to show with a snippet to appear as a supporting link in AI Overviews or AI Mode. Its AI guide adds that your site can't be excluded from generative AI features in your Search Console settings.

Outside Google, the big AI answer engines have their own crawlers (the bots that read websites). OpenAI says sites that block OAI-SearchBot won't be shown in ChatGPT search answers. Perplexity recommends allowing PerplexityBot if you want to appear in its results. So before I write anything for a site, I check its robots.txt file, the small file that tells bots where they can and can't go.

Being worth quoting

Google's AI guide draws a line I love: commodity versus non-commodity content. Commodity content is the generic stuff anyone could write, like a list of tips you've already seen fifty times. Non-commodity content brings first-hand experience, a real point of view, or details that only come from doing the work.

That's why I put my own observations, real examples and honest opinions into every piece. A summary of other people's summaries gives an AI system nothing new to cite, and gives your reader nothing new to remember.

👉 To show up in AI answers, be easy to reach and worth quoting. Most of the rest is noise.

Change #4: I stopped counting keywords

I still do keyword research, because I need to know what people type and what they mean by it. What I stopped doing is counting how many times a phrase shows up on the page.

There are two reasons. First, keyword stuffing (filling a page with keywords to manipulate rankings) is listed in Google's spam policies. I also flag it as one of the SEO mistakes that quietly hurt businesses. Second, Google's AI guide says its systems understand synonyms and general meaning, so you don't need to cram in every long-tail variation of a phrase.

So I write the way your customers talk. If your reader would say "my site is slow," I won't force "website speed optimization services" into every paragraph. The page still covers the topic fully. It just sounds like a person wrote it.

This isn't new for me, either. When I revived a dead Amazon affiliate site, the content overhaul was about value over volume: better articles, not more of them. The site saw 583% growth in clicks in three months.

Try it yourself: how I write content, one page at a time

You don't need to rewrite your whole site. Start with one page that matters, like your main service page or your most visited post, and work through these:

1. Move the answer up. Rewrite the first sentence under each heading so it answers that heading.
2. Check every claim. Find the source for anything about how Google works. If you can't find one, soften the wording or cut it.
3. Add something only you know. A real example from your business, a mistake you see customers make, or a number from your own work.
4. Read it out loud. If you'd never say a sentence to a customer, rewrite it. This catches keyword stuffing fast.
5. Break up the walls of text. Short paragraphs, clear headings, and a list or table where it helps. It's one of the content strategy mistakes I see most.
6. Make sure bots can reach it. Check that robots.txt isn't blocking Googlebot, OAI-SearchBot or PerplexityBot.

Then do it again with the next page. Refreshing what you already have is often faster than writing something new, and I explain why in my post on updating your website content.

👉 One genuinely better page beats ten new pages that say what everyone else says.

People also ask

Is SEO content still worth writing now that AI Overviews exist?

Yes. Google says its AI features are built on the same ranking and quality systems as regular search, and they link to the pages they draw from. Fewer people may click on some searches, so the pages that do get cited and clicked need to be the clearest, most useful ones on the topic.

Do I need to write differently for ChatGPT and Google AI Overviews?

Not really. Google says you don't need a special writing style for AI search. What helps is a clear structure, direct answers, original insight, and making sure crawlers can reach your pages, including Googlebot, OAI-SearchBot for ChatGPT search, and PerplexityBot for Perplexity.

Does Google have a preferred word count for blog posts?

No. Google's helpful content guidance says plainly that it doesn't have one, and lists writing to a word count as a warning sign. Write as long as the topic needs to fully answer your reader's question, then stop. A tight, useful answer beats a padded one.

What is non-commodity content?

It's content that goes beyond common knowledge. Google uses the term for pages built on first-hand experience, expert insight or a unique point of view, compared with generic content anyone could write. Real examples, original data and honest opinions are what make content non-commodity.

How can I see if my content shows up in Google's AI features?

Google's AI guide points to the Generative AI performance report in Search Console. It shows how your content performs in generative AI features on Google Search and Discover. If you haven't verified your site in Search Console yet, that's the first step.