Utilizing AI search for Blogs and Articles

Business Background 

Let’s start from the difference between product search in e-commerce and content search. The main difference is how the source of product or content information is built. There are seven obligatory pieces of information used to describe every product: 

• Product ID
• Title
• Information about availability
• Product type (category)
• Price
• Link to the product
• Image of the product

Of course, there are many more attributes which can describe such products. They are not obligatory but can enhance search, and the more information, the better.  
In case of content search the only obligatory attribute is ITEM ID.   

In this article, we will show you content search using our own example, because we use this search function on a daily basis, in three main places: 

• in our help center
• in our blog
• in our API documentation

Requirements 

At the beginning, to get content from those subpages, we have prepared dedicated scraper. You should gather all content from those subpages and pay attention to every element of the content which can be used later in your search.  

Based on collected content, we are able to prepare catalogs in Synerise. You can basically add any information needed, the one obligatory piece of data is Item ID. Below, you can see an example of such a catalog. It may look different depending on where the search is used and which elements you want to add. 

How to do it  - Search Configuration 

When we implement catalogs, it is time to go to search and prepare the configuration.
If you would like to add a new one, you have to do it in following way: 

• Go to AI Search
• Add index
• Choose the Catalog which you are interested in, based on which the search will work
• Add some basic description

Then you will be able to create a new search - where the most important part will be to prepare ranking of attributes collected in the catalogs. Let’s take a look at three searches we use in Synerise. 

Synerise Blog  

Below, you will see information which we have gathered from our blog. 

The information categories here are image and description. We have also link to this article and of course we have an Item ID, which is obligatory. Additionally, we have title, author surname and name. That information is collected from our website and using this we can prepare an AI search. 

On the website we can see an autocomplete section which automatically shows the section a user might be interested in. So, in this case after entering “artificial intelligence”, we can see the articles from that specific category. 

But we have also added the possibility to search by the author name or surname and the title of the article, also with the autocomplete option. 

Synerise Developers

In the case of the developer section, we have some additional information that is a little different than what we saw before. 

We have here category, description, Item ID and a link as well. But additionally, we have attribute “method” which is information if somebody want to send something to the Synerise via API or get some information from them. We have also information about permissions for each method and additional information like service, tags and a title which describes those methods.  

Using our search on the website we can search by methods so e.g. we can enter “post”. Then, we will get only those methods which can send some data to Synerise. This was possible because we got the information about the method type and because we added it to search configuration. Do not forget that we can also search by title, description and category of every post. 

Synerise Help Center

In the case of our Help Center, we also have things like category, link, ID, title and description.  

But one additional feature here is a table of contents of each post. So, what you can find in each article is also described here. 

When we search at help.synerise.com for product feed we can get two results which contain product feed both in the title and in the product category. And as you can see here those are the most important attributes in our set up section by which we can search.  

The third result does not have a product feed in the category or in the title, but it is added as one of the headers in the article in the table of contents.  That's why this result came up in the search. 

Summary

These three searches are built in one business profile. Each refers to different section of our website. Each of these sections also has a different configuration, because they have different attributes which describe the content. Based on that you can have any content and search you would like to have. Also, this tool lets you make this search as personalized as you want using our query rules. Using this method, you can decide that when someone enters a query, you can replace it with a different phrase or also synonym and show him a specific article. 

Your imagination is the only limitation, learn more about how to implement Synerise AI Search step by step and try it on your website. >>