Why does the AI powered search always return some results even for random searches that are seemingly not directly related to my content?
The AI optimized (powered) search results are based on the relevance to your content from the provided search term(s). The search relevancy accounts for misspellings, similar words with the same meaning and the “tone” of the question. This is what is known as “semantic search”.
Using semantic search, if the user provides a search for something that is even remotely related to your site content or has an association with your content, the engine will return to the most relevant items first and then return less relevant information afterwards up to the max number of permitted records.
Since the relevance engine will find some type of data relevance even to non-sensical search terms such as a long string of numbers or garbled text, it will return results regardless of if the user input contains actual words of phrases.
This relevance ranking, even for seemingly random text searches, allows the search to be incredibly accurate with SKU and product ID searches since it is always comparing similarity so if the user enters ta SKU number that is off by a digit or contains an incorrect prefix or suffix, the AI engine will still return the correct result.
Even though the AI powered search will always return some type of relevancy-based resource links, it does return incredibly accurate results that not only take into account the words or characters in the search but the actual meaning of the search even when user does not know the correct spelling or what exactly the product or service is called.
Note: Unlike ChatGPT, Shop Bot Pro will NOT return results for content outside of your indexed content even if it can’t find very relevant results that closely match the user’s query.
Instead, Shop Bot Pro will return the most relevant content resources, contained inside your indexed data, in relation to the user’s search even if the relevancy is very limited.
As mentioned above, always getting some relevant content for a search is very helpful for ecommerce sites where customers are searching by SKU, product ID, Manufacturer Number, etc since those types of queries are typically random characters which would not be returned by a standard keyword search if not typed in correctly (exactly). However, with semantic based AI results, the user will get the most relevant results even for misspelled or mis-entered SKU and product code based searches.
