Leave a request
Hello friends! Today I want to share my thoughts on how artificial intelligence (AI) affects our field. If you're not already aware, AI is the technology that we're promised will soon take over most tasks, including content creation. But is this true? Let's figure out how things are going in SEO.
Recently, our team has been talking a lot about generating descriptions for products on various platforms using AI. We decided to test several such services and compare their results. But before that, it is important to understand how we usually write such descriptions.
Here's how we usually approach writing texts:
1. We form a dictionary that includes all the necessary keywords - both high-frequency (HF) and low-frequency (LF).
2. Next, we begin to work with this list, entering all these keys into the texts in order to ultimately obtain content that contains all the selected words.
So, now let's move on to testing AI services:
Unlike humans, AI does not always keep track of all keywords. He chooses what will be included in the text and what will not.
For the experiment, we used the same dictionary of 137 words for the product “women’s dress.” Here's what happened:
1. The best work was shown by one of the popular AI services. We gave him an exact request: “write a description for the product “women’s black dress” with all the words included in full.” Result: out of 137 words, he used 89, and the missed frequency was 1,466,383, which is the lowest.
2. We got the worst result from another service. This service only used 46 words from the list, and the missed frequency was 16,324,687.
3. In order to get even more information, we tested the service on another request: “write a description for a dress in women’s size 42.” This time the result was even worse - out of 137 words, the AI used only 51, and the missed frequency was as much as 46,705,347.
What follows from this?
This does not mean that the services do not work. They create text, but not always in full accordance with our requirements. The AI generates a description using the selected words, but does not always adhere to all key phrases, resulting in significant omissions.
And the most interesting thing is that we cannot know exactly which words were used by the AI, and in what quantities. This means that in order to obtain high-quality SEO content, we still need to do reconciliation.
Once we have found the words that were missing, we have two options:
1) Insert the missing keys manually.
2) Use AI again to generate new texts and combine them.
One way or another, we will still have to manually correct the text, since even the most advanced AI systems sometimes create imperfect phrases and phrases.
If we talk about our team, we tried to introduce AI into our processes, but it didn’t work. AI does a good job of generating text, but when you need to strictly follow a list of keywords, you have to make a lot of additional edits. In some cases, it is easier and faster to write text from scratch.
Have you tried using AI for SEO? What experience do you have? Share your impressions!
If you have any questions about SEO promotion, you can contact the SEO studio "SEO COMPUTER" for any question by email info@seo.computer.
ID 4908
