HubSpot’s 2024 State of Marketing Report found that marketers save around 2.5 hours per day with AI. Further, 60% of marketers see AI tools as helpful assistants in their jobs.
When it comes to technical SEO and AI, specifically, I set out to answer three burning questions: Are SEO folks using AI? If so, how? And is it actually worth it?
To find answers, I contacted members of HubSpot’s technical SEO team and practitioners from my external network.
If you’re asking yourself the same questions, you’ve come to the right place. Let’s get into the good stuff!
Table of Contents
How to Use AI for Technical SEO
Adding AI to Your Technical SEO Strategy: Yay or Nay?
How to Use AI for Technical SEO
Technical SEO refers to anything you do that makes your site easier for search engines to crawl and index. Technical SEO, content strategy, and link-building strategies all work together to help your pages rank highly in search.
AI can be leveraged in various ways to help your technical SEO strategy. The first five use cases are from HubSpot’s internal SEO team.
The final three are from SEO practitioners in my network doing interesting things with AI for tech SEO purposes.
1. Improve Internal Linking Architecture
I know what you’re thinking: Isn’t internal linking on-page SEO? As with most things in SEO, it depends. (You knew that was coming at some point, folks, so I got it in early for good measure.)
You might not figure out anchor text for individual links as part of your technical SEO efforts. However, you will need to factor in internal linking as part of a site’s overall link architecture. To get a clear picture, you’ll need a bird’s-eye view of any existing internal linking.
Enter Killian Kelly’s use case for AI. Kelly is a marketing manager and SEO content strategist working on HubSpot’s EN blog strategy.
“I‘m currently looking into ways we can enhance our blog’s internal linking structure by using Screaming Frog‘s Ngram tool,” says Kelly.
“It’s going to be incredibly useful for navigating through the large amount of content on HubSpot’s blog. The tool is excellent at identifying both linked and unlinked keywords within the content, which is very helpful for internal linking.”
Kelly adds, “During this exploration, I stumbled upon the possibility of integrating OpenAI with Screaming Frog. This will allow us to use ChatGPT prompts during the crawl, which can help automate and optimize different aspects of SEO.”
Kelly explains that the integration should help automate tasks like writing and optimizing alt text for images, creating anchor text, and structuring data automatically.
It can also automatically classify the content into themes, which he thinks would be very helpful for understanding page intent and possible correlations between content types and performance.
Best for: Many of the AI automations mentioned above are more relevant to on-page SEO. So, I think this use case would benefit you if you have a mixture of on-page and tech in your role. And let’s face it, that’s a lot of us these days.
2. Generate Schema Tags
The next four use cases come from Sylvain Charbit, the senior marketing manager on HubSpot’s tech SEO team. He discusses using AI to generate schema tags, conduct log file analysis, and more.
“There are a few ways to use AI for technical SEO, the most common one being to generate Schema tags,” says Charbit. “These small blocks of code are used to display rich results in search (among other things).”
There are a ton of AI-powered schema markup generators available online. If you’re already using Jasper.ai for your content efforts, you can use Jasper Chat to create schema markup.
Of course, there’s OpenAI’s ChatGPT. I know of many folks who use the free version for this task.
As with anything AI-related, you’ll want to validate what the tool spits out to make sure it’s accurate and functional.
Best for: SEO practitioners who aren’t very comfortable with coding. But even if you are, it can help you save time.
3. Log File Analysis
“Another way we leverage AI for technical SEO is to have it analyze part of our logs (the one without sensitive user data) and recognize behavior patterns,” says Charbit.
“Maybe Googlebot is getting stuck somewhere or crawling many URLs with no interest. AI allows us to know what is going on in a flash and to act accordingly.”
Disclaimer alert: I haven’t personally tested this. However, you should be able to use the Data Analysis GPT in ChatGPT to add some AI magic to your log file analysis. Theoretically, it should be as simple as dragging and dropping your log file into the GPT and asking it a question to get started.
If you’d like to explore this concept further, I found this tutorial pretty helpful!
<iframe width=“560” height=“315” src=“https://www.youtube.com/embed/8CpXdAamxeA?si=Gu_8IY8MuUpssmZX” title=“YouTube video player” frameborder=“0” allow=“accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share” referrerpolicy=“strict-origin-when-cross-origin” allowfullscreen></iframe>
Best for: SEO practitioners who want to recognize behavior patterns at speed.
4. Get a Second Opinion on Your Code
HubSpot’s Sylvain Charbit shares another use case for AI in technical SEO.
“Additionally, getting an opinion from AI on a block of code can be useful to detect an issue if a manual review doesn’t provide any results,” Charbit says.
I’d say that in this case, whatever AI tool you use to validate your code, it should, as the cool kids say, function as your “intern.”
That’s opposed to taking the lead with your code. In short, for this use case, you must have the skills and knowledge to recognize whether AI is hallucinating.
Best for: SEO practitioners with a competent understanding of code but would like a second opinion.
5. Communicate Technical Ideas to Decision-Makers
“Last but not least, communication! Being able to simplify and communicate technical SEO is crucial to getting buy-in from decision-makers,” says Charbit.
He adds, “As I am constantly head down into technical stuff, I can sometimes forget that some terms or facts are not known to many people. AI reminds me of this and improves collaboration with multiple stakeholders by providing more digestible information.”
If I were to do this, I think I’d probably enlist Grammarly, one of my go-to AI-enhanced tools, for the task. Why? It’ll accomplish two things at once.
First things first, it’ll check the text for spelling, grammar, and tone. Second, you can use the new “Increase the impact of your text” feature to highlight any parts of the text that you might need to clarify for a beginner audience.
I’d also highly recommend the “Clarity” function. I’ve been really impressed with how much that’s improved in the last year or so!
I’d start by setting the “Audience” goal to “General.” I’d then click “Increase the impact of your text.”
I do have a word of warning for this Grammarly feature, though. Sometimes, it can be a little over the top with its suggestions. Then, before you know it, rather than filling in your audience’s potential knowledge gaps, you’re actually talking down to them.
I generally use this feature to highlight potential areas for improvement rather than using the solutions presented by Grammarly.
Best for: SEO practitioners who need to communicate technical ideas — including the business benefits of implementation — to non-technical stakeholders.
6. Track Headers During Audits
Next, Mike Ciffone, an SEO consultant at Ciffone Digital, shares how he’s been using AI for technical SEO audits.
“While perhaps not the most glamorous use of AI, in my audits, I’ve been using it to keep track of headers,” says Ciffone. “When I fire up Screaming Frog, I use the JavaScript execution feature to store the HTTP response of each URL. Then, with AI, I’m simply asking questions and getting told the story.”
Ciffone asks AI questions like:
Are there any patterns in response codes?
Where are we getting the most cache misses?
Do I have X-robots headers setting no-index/nofollow or canonicals anywhere?
He adds, “In my opinion, there’s way too much instinct involved in auditing for AI to be very useful for any sort of automation (for now at least). However, as a personal assistant, it’s drastically improved my efficiency and shortened my turnaround times.”
Best for: SEO practitioners working with the combination of having a separate mobile site (e.g., m.example.com), multiple language and geo versions, and also working with a progressive web app versus merely a responsive site.
7. Deploy Schema at Scale
I wanted to build upon the schema tag generation use case presented by Sylvain Charbit (number two on this list.)
So I contacted International SEO Consultant Aarne Salminen, who I noticed talking about generating schema templates in bulk to deploy sitewide in MostlyMarketing’s Slack community.
“I do this for sites that have hundreds of content types = schema templates and millions of URLs,” says Salminen. “If you have just a few types of content, I might not go the AI route, but on large-scale projects, it seems to speed up things, including setting everything up.”
Salminen adds, “I don’t use AI in any active component in the process because reliability is most likely still an issue. So it is in the preprocessing stage and/or planning stage, where you build up templates per content type, keeping the big picture of the website infrastructure and internal connections in mind.”
I asked Aarne to share what this process looks like. He said he feeds it the data of their site, such as Screaming Frog type of data with identified and manually verified page types, and lets the AI run the first pass of suggestions.
After that, he verifies and validates the AI input, tweaks it, and does a second pass if need be. Then, he verifies and validates again, and finally, it goes to implementation.
Best for: SEO practitioners working on sites that have large quantities of content types and URLs.
8. Visualize Google Search Console (GSC) Data
Last but not least, I learned about this use case from Sreeram Sharma, an SEO consultant and co-founder of Angleout.
“I use ChatGPT to visualize the GSC data while looking for pages that were hit or gained traffic during a specific time period,” says Sharma. “This helps me to plot a graph and visualize stuff rather than using Tables or Looker Studio. I like using this approach compared to Looker Studio.”
I asked Sharma to expand upon the process:
“I run a screaming frog audit and export it into sheets, then use vlookup to map them with clicks + impressions before/after traffic drop. Now, I upload the sheet to ChatGPT and ask it to visualize and show a correlation of the number of tech errors versus the drop in traffic.”
Sharma adds, “This helps me get an approximate idea of traffic drop and makes it a bit easier to explain to my clients on monthly calls. So far, they’ve loved it.”
Best for: SEO practitioners looking for an alternative to Tables or Looker Studio.
Adding AI to Your Technical SEO Strategy: Yay or Nay?
Ever since OpenAI unleashed ChatGPT in the winter of 2022, there’s been a ton of hype around AI.
Upon the back of the release — and seemingly in the blink of an eye — we went from AI being a developing concept, bubbling away in the background, to it being everywhere. Integrated into anything and everything.
The floodgates had truly opened.
Now, it‘s the summer of 2024. You look to your left: AI. You look to your right: Oh, hello, that’s some more AI. But unlike 2022, the dust has settled somewhat, and maybe you’re like me, constantly asking: Is the juice really worth the squeeze?
When it comes to AI for technical SEO, the answer is both yay and nay. Honestly, it depends on your unique situation. That said, there are two things I can say with absolute confidence:
1) There‘s literally no point using AI for tech SEO simply for the sake of it. If it doesn’t add value to your process (i.e., save you time and improve efficiency), it’s hype — plain and simple.
2) If you remove the human from tech SEO at this stage, where AI is right now, you’re cooked.
On the latter point, will this change in the future? Who knows. I personally don‘t think you can ever fully remove humans from SEO. But that’s just my humble opinion.
What do you think?
