Can AI perform technical SEO analysis from flat data?

Generative AI has made important strides in varied fields, together with SEO.

This text explores a vital query: Can AI successfully carry out technical search engine optimisation evaluation utilizing uncooked crawl information?

We’ll study the capabilities of ChatGPT in decoding Screaming Frog crawl information and producing search engine optimisation suggestions. 

Leveraging generative AI for technical search engine optimisation

My Search Engine Land articles usually study in-depth technical purposes the place search engine optimisation and AI can collaborate, leading to elevated efficiencies. For instance, I’ve used AI to generate alt text from image files or product descriptions from PIM data.

Nevertheless, many entrepreneurs are much less technical and will battle to mix OpenAI’s API, Python and spreadsheets. Some entrepreneurs would fairly hand AI the tech information and obtain clear human insights. 

As such, I believed I might produce an article analyzing the effectiveness of AI in decoding technical search engine optimisation information. Even for those who’re not speaking with AI fashions through programmatic API entry, many AI assistants (ChatGPT, Google Gemini) provide user-friendly chat interfaces. We are going to discover ChatGPT on this article.

Recently, the power of those AI-powered chat interfaces has been rising. For instance, OpenAI’s ChatGPT has lately advanced from GPT-4 to GPT-4o. The brand new GPT-4o mannequin comes with many function modifications:

• GPT-4o is way sooner than GPT-4. It has the same customary of accuracy to GPT-4, whereas responding at tempo, extra just like GPT-3.5-Turbo.
• GPT-4o is much less reliant on plugins or customized GPTs to entry internet content material.
• GPT-4o will search internet content material much more continuously to generate its outcomes, making the produced materials “more energizing.”
• Although true, the net is poorly curated and stuffed with misinformation. Many really feel that the refined information mannequin of GPT-4, with fewer internet requests – really produced superior outcomes to GPT-4o.
• There’s no arguing that GPT-4o is a minimum of corresponding to GPT-4. It’s a lot sooner and much more interactive. It could work with extra kinds of connected information, produce extra kinds of media and is web-capable.

With GPT-4o now extra able to looking internet content material, it could possibly be extra useful for entrepreneurs, even those that lack technical abilities. Nevertheless, there are issues about its potential to entry the net, as previous experiences with instruments like Google’s Bard and Gemini confirmed inferior outcomes.

To check GPT-4o, we’ll give it particular technical search engine optimisation crawl information as a substitute of generic evaluation duties. This may assist us see how a lot ChatGPT has improved and if it may be helpful for technical search engine optimisation evaluation.

Analyzing Screaming Frog crawl information for insights through ChatGPT

The “inside all” export from Screaming Frog is the bread-and-butter basis of most technical search engine optimisation insights. 

That single export could also be pivoted in a number of methods to touch upon metadata points, canonical tag conflicts, hreflang points and so on. 

Will AI decide up on such points? Will AI hallucinate or misdiagnose points the place none are current? Let’s discover out.

We are going to use the Butcher’s Canine Meals firm web site as a testbed for these exams. They’re a UK pet food provider, and their web site is about the precise measurement for our proposed actions.

First, we are going to crawl Butchersdogfood.co.uk:

Over 3,000 inside addresses had been found. Now, we will export the record of inside addresses to a CSV file:

Now, we will feed the CSV file to GPT-4o and ask for some suggestions. Most individuals don’t write prompts as complicated as I do. They write shorter prompts after which refine them. As such, I’ll try to comply with that habits.

The response contained many suggestions, however most had been disappointingly generic. For instance, suggestions like this:

• “A number of URLs level to picture recordsdata (e.g., PNG, JPEG). Be sure that all photographs are optimized for internet utilization to scale back load occasions. This consists of compressing photographs and utilizing applicable codecs.”

This suggestion is disappointing as a result of ChatGPT really did have entry to response time and file-size information as per picture addresses. So, the response may doubtlessly have been extra particular.

Right here’s one other instance of very generic steering:

• “Be sure that canonical tags are appropriately carried out to keep away from duplicate content material points. Every web page ought to have a self-referencing canonical tag until it’s a deliberate duplicate.”

Since ChatGPT has entry to the datasheet, it ought to have the ability to work out (roughly) which URLs are the primary addresses and which aren’t (parameters within the URL string, and so on.).

Thus, ChatGPT ought to have the ability to isolate canonical tag issues fairly than give common (and ineffective) steering.

Let’s push it a bit more durable:

The steering remains to be comparatively obscure, even once we iterate and create a extra detailed immediate. Many pages are all the time lacking canonical tags, corresponding to picture addresses, which don’t help HTML.

The response ought to have been extra particular and detailed, explaining why sure addresses wanted canonical tags however didn’t have them.

As a part of the response above, some extra detailed suggestions was given:

Wow, these suggestions are a complete mess.

In a method, they’re spectacular regardless of being completely fallacious. I’d anticipate an inexperienced junior search engine optimisation particular person whose been thrown into the technical search engine optimisation deep finish to make comparable suggestions. 

The recommendation could be fallacious, however there are certainly indicators of intelligence. You have to keep in mind that it nonetheless takes a point of intelligence to go off on tangents and produce related but inaccurate evaluation. The work is unusable, however there are indicators of a real try.

As people, we’re susceptible to holding AI to unfair requirements. If AI produces one thing that’s fallacious or incorrect, we are saying that AI has failed to attain human understanding. Nevertheless, many people (actual people) additionally fall brief, and this occurs usually. 

For me, the query is: Has AI failed in a method that demonstrates some human-like try at finishing the duty?

I believe indicators are optimistic for the longer term when it comes to such purposes for accessible (chat interface) AI interactions. For now, although, it will be disastrous to behave on the suggestions given. 

Figuring out a redirecting URL as doubtlessly necessary (although it’s a redirecting deal with) after which assuming that work might or might not have to happen on the ultimate vacation spot URL is a few of the most woolly pondering I’ve personally encountered.

The one attention-grabbing pointer that ChatGPT was in a position to provide you with was that plenty of addresses on the location resulted in 301 redirects, and there could also be some faulty structure:

The above response was given after our very first message earlier than we started pushing AI to assume in additional particular phrases. This was really the primary suggestion that ChatGPT gave, and in some methods it’s attention-grabbing:

There are round 100 redirecting addresses. Many of those are redirected picture addresses. Some look like they had been as soon as true pages or maybe are trailing slash redirects. Others revolve round pagination.

This was an attention-grabbing lead, however most different suggestions had been generic or deceptive. ChatGPT’s try at particular recommendation was semi-intelligent, however the solutions would have wasted time or precipitated extra issues. 

I’d grade this work a D or E, although many machines wouldn’t rating that top on technical search engine optimisation from flat datasheets. I’m impressed, however I don’t suggest utilizing AI for this degree of technical search engine optimisation evaluation but.

Evaluating AI insights to Ahrefs’ technical audit

Ahrefs has a complicated cloud-based search engine optimisation crawler. Whereas Screaming Frog’s information might be extra correct, Ahrefs nonetheless supplies stable information and higher insights. I exploit each instruments for web site audits.

If we have a look at the output from a full Ahrefs technical search engine optimisation crawl, we might then have the ability to decide how shut GPT-4o got here to discovering the reality. Let’s have a look at the Ahrefs overview:

Wow, that’s a superb ranking! Effectively carried out to the Butcher’s Canine Meals workforce.

Regardless of the excessive ranking, some points nonetheless persist. That is completely regular and is often nothing to fret about. 

What’s attention-grabbing right here is that the remaining points revolve round redirecting addresses and pictures, which can have too massive file sizes. That is attention-grabbing as a result of these had been a few of the prime points picked out by our AI assistant primarily based on Screaming Frog information.

What does this imply? Was the AI proper all alongside, and maybe we judged ChatGPT too harshly for “generic sounding” suggestions that had been really correct? 

Sure and no. People usually really feel the necessity to defend their roles towards advancing know-how, which may create a bias towards it. This bias is frequent, even amongst these within the digital subject.

Nevertheless, this know-how shouldn’t be a direct risk to technical search engine optimisation specialists. The recommendation given was not usable, however the AI’s errors had been clever, just like these of an inexperienced human who exhibits promise. 

The final focus of the insights, like redirects and picture compression, was correct even when the precise recommendation wasn’t.

The ultimate verdict

GPT-4o is an enormous enchancment over GPT-3.5-Turbo. Nevertheless, I’m unconvinced that there’s important (or any) enchancment over GPT-4, which appears to hallucinate much less. 

Presently, I see GPT-4o as a very good center floor between GPT-3.5-Turbo and GPT-4. To me, the outcomes of GPT-4 appear superior. However GPT-4o is round as quick as GPT-3.5-Turbo and is extra interactive. 

I suppose time will inform which mannequin customers will favor. I’m keen to attend longer for GPT-4 to provide superior output, even when that output is restricted through a (curated) information mannequin.

ChatGPT’s recommendation was poor. It missed key factors and provided generic, unhelpful solutions. Its try at particular recommendation additionally failed.

Nevertheless, there have been indicators of intelligence in its errors, just like an inexperienced human making errors. This exhibits potential for enchancment.

For now, don’t depend on this know-how for technical search engine optimisation evaluation, particularly by way of easy chat interfaces. However control it – an AI able to offering invaluable insights could also be nearer than we expect.