Feeding data to Azure Open AI

Hi all,

I’m switching from the standard Open AI API to Azure, I have to say it’s great so far.

I am trying to set it up such that my users can ask the AI questions about the data within the app, for example, “Summarise the engagement from the last 2 weeks” (Where engagement is a data type containing things like emails, notes from calls etc).

I know this is possible but I’m struggling to find my way - does anybody have any suggestions here please?

While there might be ways to achieve this now, Bubble is already working on incorporating AI into the product. I imagine what you are trying to achieve will be the starting point for Bubble’s AI feature. I’m sure more on this will be shared at next month’s Bubble conference.

Interesting, I didn’t know they were working on this. Do you have any further details/link?

Do a search for Engagement (Created Date > Current date/time +weeks -2):format as text

Display each engagement in the prompt to OpenAI and say ‘using this data… do X’.

It’s been mentioned by Josh several times, one of them is here Monthly Community Update -- August 2023

1 Like

Thanks - this gets me somewhere near.

I’m also trying to do a similar thing but with a more complex data set, in this case, I have a data type “RAG Rating” with “Notes”, “Categories”, “Reasons” and “Actions”. I am aiming for the AI to be able to look at all of this data and roll it up into a conversational summary.

Any ideas?

If you provided the database schema in the prompt and some info on formatting Bubble Data API requests you could get ChatGPT to generate a Bubble Data API request that you can use to get the data and insert the data into another prompt. Kind of complex though and doesn’t really allow you to structure the resulting data in the prompt.

2 Likes

Ok thanks this is helpful