AI Studio, How can i update data without creating an entirely new vector index?

Middendorf, Jelle (ZGNL) Greetings & Welcome to Q&A forum!

How can I update or add to the data my vector index uses so that my chatbot can access additional information without having to recreate it every time from scratch and change the reference.

Did you check this build and consume vector indexes in Azure AI Studio already? You can schedule updates. Also, you can choose to update the index hourly or daily.

Once the Indexer is established, you can also directly modify the schedule settings through the Azure AI Search Portal. See Update or rebuild an index in Azure AI Search and Schedule automatic index refreshes for more details.

To add to this, how would you do this if i deployed the chatbot to an endpoint. Is it static once deployed?

AFAIK, If you want to update the chatbot with new data, you will need to retrain the model and redeploy it to the endpoint.

Do let me know if that helps or have any other queries.

@AshokPeddakotla-MSFT Oh nice, that looks very promising.
So i went through the steps you gave me the link to (https://video2.skills-academy.com/en-us/azure/ai-studio/how-to/index-add) and as you can see at step 12 (image below) it shows the option for schedueled updates. But i still have an issue.

I have only worked with uploaded files so far so i thought if i select azure blob storage it must work for me as well. Sadly when I do I dont get shown this option. (Image below)

In your response you show where you can select it but i cant find it. It looks very similar to the image in (https://video2.skills-academy.com/en-us/azure/ai-services/openai/concepts/use-your-data?tabs=blob-storage%2Ccopilot#schedule-automatic-index-refreshes).
I have an azure blob storage so i should be able to try this out, could you be so kind to explain how you navigated to that window?

And about the deployed chatbots, My chatbot uses RAG so i dont have to train it on the data. So as far as I know the chatbot simply references the data from the vector index. So when the vector index automatically updates with new data added to the blob storage, wouldn't my deployed chatbot now be able to access the updated/new data?

You have been very helpfull so far, thank you very much.

Middendorf, Jelle (ZGNL) I'm looking into this and update you as earliest. Appreciate your time and patience.

@AshokPeddakotla-MSFT Thanks for the help so far. It’s been a few weeks, and I wanted to follow up to express that I’m still very interested in the answer. Could you tell me when I might expect a response? I really appreciate your help.

Middendorf, Jelle (ZGNL) Apologies for the delayed response. The earlier imaged referenced is in Azure OpenAI studio.

I have an azure blob storage so i should be able to try this out, could you be so kind to explain how you navigated to that window?

Please follow the below steps to get that option in Azure AI studio.

Select the Add data a new data source option in the AI studio -> Select Data Source as Blob Storage where your data is available.

Select next-> choose the search resource you have created or create a new search resource -> you can see the option to select scheduling updates

I hope this clarifies.

And about the deployed chatbots, My chatbot uses RAG so i dont have to train it on the data. So as far as I know the chatbot simply references the data from the vector index. So when the vector index automatically updates with new data added to the blob storage, wouldn't my deployed chatbot now be able to access the updated/new data?

Considering you'd configured RAG, yes the data will be updated as per the schedule.

Do let me know if there are any further queries I can help with.