I have created azure ai search service called centel**dataservice1. After this, I created skillset called text split skill. After this, I created index called azureblob-index1 using azure blob storage in which there is a container called rag where all my important documents are stored.

Then, I created chat playground using azure ai search index. Also, created web application of the same.

When I am asking question in chat playground and web application, the responses are different and references are also different.

I asked Question:

How does Cen**lon plan to ensure the accessibility and inclusivity of the Digital Curriculum Website for diverse user groups?

This is image of Web application where question is asked, got response and reference

After this, I asked same question in chat playground.
This image is of chat playground where same question is asked, got different response and got different reference

I have used fields of the index like:

content (String)
metadata_storage_content_type (String)
metadata_storage_size (Int64)
metadata_storage_last_modified (DateTimeOffset)
metadata_storage_path (String)
metadata_author (String)
metadata_title (String)
metadata_creation_date (DateTimeOffset)
language (String)
split_text (StringCollection)
keywords (String)
summary (String)
section_titles (String)
metadata_file_type (String)
merged_content (String)
text (StringCollection)
layoutText (StringCollection)
metadata_storage_file_extension (String)
metadata_content_type (String)
metadata_language (String)
metadata_storage_name(string)

There are only 5 documents in my container right now.

I have used custom_mapping like thisthis image explains what I have kept in custom_mapping

I am not able to understand why responses are different. Same question I asked and got different response. Also, the reference document which is given is wrong.

What can be the exact reason.