First, I use the method from @HenryJammes ,It's good for my prod env to use PVA ChatBot to answer Employee questions,But sometime someone input key words it not good, and will get not good answer from my bot.
Two weeks ago,I read Azure OpenAI on your data (preview)This way support semantic search.
Maybe, It's a better way for use PVA chatBot answer your Employee questions.
Here is my Step to use Azure OpenAI on your data (preview) in my PVA:
1. I follow Quickstart: Chat with Azure OpenAI models using your own data complete the setting and upload my data.
2. I Create an Topic named "Cognitive search"
3. I create Power automate flow in step 2's action
4.I trigger the flow and got output from Azure Cognitive search
{
"id": "12345678-1a2b-3c4e5f-a123-12345678abcd",
"model": "",
"created": 1684304924,
"object": "chat.completion",
"choices": [
{
"index": 0,
"messages": [
{
"role": "tool",
"content": "{\"citations\": [{\"content\": \"\\nCognitive Services are cloud-based artificial intelligence (AI) services...\", \"id\": null, \"title\": \"What is Cognitive Services\", \"filepath\": null, \"url\": null, \"metadata\": {\"chunking\": \"orignal document size=250. Scores=0.4314117431640625 and 1.72564697265625.Org Highlight count=4.\"}, \"chunk_id\": \"0\"}], \"intent\": \"[\\\"Learn about Azure Cognitive Services.\\\"]\"}",
"end_turn": false
},
{
"role": "assistant",
"content": " \nAzure Cognitive Services are cloud-based artificial intelligence (AI) services that help developers build cognitive intelligence into applications without having direct AI or data science skills or knowledge. [doc1]. Azure Machine Learning is a cloud service for accelerating and managing the machine learning project lifecycle. [doc1].",
"end_turn": true
}
]
}
]
}
not complete yet, and will update.
Hi Henry,
I have deployed my GPT432K model as a PVA in copilot/I can see a new Conversational Booster topic has been added but I can see the model details if I edit the DataSource. But whenever I am asking any question, it's going to default fallback and I am not getting any error message as well. Any idea what I might be doing wrong? I have used semantic search and also have selected standard tier while creating the index.
Once you add your data source and configure it with vector search it will automatically create an endpoint which you can use in Power Automate by using the HTTP connector. You can see the end point details by clicking on "view code" above the chat window:
hey @remidyon, I am trying to deploy to a PVA bot, so it looks like vector search is still unsupported. I am wondering if there may be a workaround you might suggest that could possibly entail a power automate flow, or HTTP request to cognitive search but using vector search rather than semantic
It seems like vector models support has been added:
@remidyon @HenryJammes would it be possible to implement this with vector search rather than semantic?
Hi @remidyon Thanks for the idea!! It works.. 🙂
@AbigailCuadra you can manage security and access to your blob storage. I am able to click on the link in PVA and open the PDF stored in blob storage after I configured it correctly.
I would use URLs that users can access, for example SharePoint ones, if documents are indexed from that source.
Hi @HenryJammes Thank you for your quick reply and letting me know that it is work in progress. So, for now, what would be the work around or suggested channel that is the suitable as of the moment where we can deploy our PVA bot to support citation?
There is work in progress to support citation management, but this will of course depend on the channel and where you deploy your PVA bot.
Henry
Pablo Roldan
27
Romain The Low-Code...
25
stampcoin
10