First Things First
Before anything else, test your website in PageSpeed Insights: https://pagespeed.web.dev/. That way you’ll know roughly where your site sits and what the issues might be. You’ll get a performance score for mobile and desktop, as well as a breakdown of things that can improve the score.
Why it Matters
Usability - Humans First
It’s easy to buy into the idea that people are going to stop using your website entirely in favor of AI snippets. It’s an idea that big tech are selling and I have some strong opinions on that. But how quickly your site loads on desktop and mobile is still a usability issue. People don’t want to wait 12s on mobile for your site to load. This effect compounds if you have an ecommerce site (affecting browsing, product comparison, cart, checkout, etc.).
Real case: I benchmarked and optimized a site recently that had a 14s speed index on mobile. 7s of that was a YouTube embed on their homepage. Adding an image that links to a video rather than a third-party embed is not only good for speed, but also for privacy and GDPR compliance.
SEO - Especially for Ecommerce
It’s true that speed and SEO in general have taken a huge hit during the AI revolution. Google still uses speed in its Core Web Vitals (LCP, INP, and CLS are still used) so I think it’s a sure bet that Google is also using those scores when deciding who to scrape for training material.
One area where speed is still critical in SEO is in ecommerce solutions, especially with product structured data (Google Shopping), so there are a lot of reasons to make sure your store is scoring well.
GEO - Optimizing for AI
We could get into an interesting discussion about AI ethics, training material and sustainability, but there’s a lot of that online already. So let’s focus on speed.
If you want your site to appear in AI-based results (chat, snippets, etc.), you need your site to be easily accessible to their training bots.
Speed affects how easily your site is scraped by bots. Do you think AI companies are going to waste their time and money on slow sites?
Let’s consider:
1. They send their indexers to a URL to harvest your content.
2. If your URL responds slowly, those indexers are waiting for the payload.
3. The indexers parse your content, possibly looking for further links to add to their data.
4. Rinse and repeat until they have what they want.
With that in mind, it makes sense for indexers to prioritize faster and easier sites to scrape because otherwise it’s costing them time and money. I’d like to add that this has always been the case. Google indexers always followed that same process. The difference now is the amount of content ingested and the way it's used.
Structured Data
They are likely trying to get the maximum amount of milk with a minimum amount of moo, by using data endpoints and any other structured data that they can parse quickly. Why would they go to the main site when they can access your endpoint just as easily?
Half-Pages
They likely prioritize the first half of a page so that’s where the main content should be (and it should load quickly). It would make sense that they only take part of a page and skip the ‘fluff’ that humans like and need, such as footers.
JavaScript Content
Another thing to consider is content loaded by JavaScript.
Indexing sites that don’t have content in the actual DOM/HTML takes more resources since the whole page has to load first. It’s one of the reasons my newest site has a Laravel backend that builds static HTML through Astro. I want the speed and benefits of proper caching, and I want bots to easily see content.
This is something to consider when vibe coding as well, as agentic coding often prefers front-end-heavy stacks which can leave bots with little to actually index.
Hopefully this has given you some insight into your own website and some ideas on how to optimize it whether for people, bots, or both!
