A RAG chat app that answers natural-language questions about rows in a PostgreSQL table, using pgvector search and OpenAI models, deployable to Azure.
microsoft/Conversation-Knowledge-Mining-Solution-Accelerator
Azure-based solution accelerator that mines conversational and document data with generative AI, offering chat exploration and auto-generated insight dashboards.
- Stars
- 478
- Forks
- 286
- Open issues
- 16
- Last commit
- 3 days ago
What is it?
Conversation Knowledge Mining is a Microsoft solution accelerator for analyzing large volumes of unstructured conversational data such as call transcripts, PDFs, Word documents, and audio recordings. Uploaded or connected data passes through an extraction pipeline built on Azure Content Understanding and Azure OpenAI, which pulls out key phrases, topics, and entities, then generates embeddings and indexes the results in Azure AI Search.
Users explore the data through a chat interface backed by a Microsoft Foundry agent that queries both semantic search and a SQL database, returning grounded answers. A separate Insights view has the LLM read the dataset schema and plan KPIs and charts, so the dashboard layout adapts to each dataset. Three built-in sample scenarios cover IT helpdesk calls, mortgage documents, and telecom transcripts, and existing Azure AI Search indexes or Microsoft Fabric sources can be connected without re-uploading.
Deployment is done with the Azure Developer CLI (azd) into your own Azure subscription, and the project uses managed identities rather than hard-coded credentials. A production-oriented configuration adds private networking and private endpoints.
Key features
- 01Extracts key phrases, topics, and entities from transcripts, documents, and audio.
- 02Chat interface that answers questions by combining semantic search with SQL analytics.
- 03LLM-planned dashboards whose KPIs and charts are generated from each dataset's schema.
- 04Accepts PDF, DOCX, JSON, CSV, TXT, image, WAV, and MP3 files.
- 05Connects existing Azure AI Search indexes and Microsoft Fabric data sources.
- 06Deploys via Azure Developer CLI with managed identity authentication.
Use cases
- Mining patterns and sentiment from IT helpdesk call transcripts.
- Summarizing and reviewing mortgage housing reports and purchase contracts.
- Analyzing telecom call recordings and transcripts at scale.
- Adding a conversational interface over data already stored in Azure AI Search or Microsoft Fabric.
Quick start
- 1
Install Azure Developer CLI (azd) version 1.18.0 or higher
- 2
Check Azure OpenAI quota by following the quota check guide
- 3
Follow the quick deploy steps in the deployment guide to deploy to your Azure subscription
Things to know
- Requires an Azure subscription and runs on paid Azure services, so costs are usage-based.
- Needs Azure OpenAI quota and a region where the required Azure services are available.
- Azure Container Registry charges a fixed daily cost even when the app is idle.
Topics
FAQ
- What is Conversation-Knowledge-Mining-Solution-Accelerator?
- Azure-based solution accelerator that mines conversational and document data with generative AI, offering chat exploration and auto-generated insight dashboards. Conversation Knowledge Mining is a Microsoft solution accelerator for analyzing large volumes of unstructured conversational data such as call transcripts, PDFs, Word documents, and audio recordings. Uploaded or connected data passes through an extraction pipeline built on Azure Content Understanding and Azure OpenAI, which pulls out key phrases, topics, and entities, then generates embeddings and indexes the results in Azure AI Search.
- Is Conversation-Knowledge-Mining-Solution-Accelerator free to use? What license does it have?
- Conversation-Knowledge-Mining-Solution-Accelerator is open source under the MIT license. Check the license terms before using it in commercial products.
- Is Conversation-Knowledge-Mining-Solution-Accelerator still maintained?
- Actively maintained. The last commit to Conversation-Knowledge-Mining-Solution-Accelerator was on Nov 4, 2026.
- What can I use Conversation-Knowledge-Mining-Solution-Accelerator for?
- Mining patterns and sentiment from IT helpdesk call transcripts. Summarizing and reviewing mortgage housing reports and purchase contracts. Analyzing telecom call recordings and transcripts at scale. Adding a conversational interface over data already stored in Azure AI Search or Microsoft Fabric.
- How do I install Conversation-Knowledge-Mining-Solution-Accelerator?
- Install Azure Developer CLI (azd) version 1.18.0 or higher Check Azure OpenAI quota by following the quota check guide Follow the quick deploy steps in the deployment guide to deploy to your Azure subscription
- Where can I get Conversation-Knowledge-Mining-Solution-Accelerator?
- The source code and installation instructions are on GitHub: https://github.com/microsoft/Conversation-Knowledge-Mining-Solution-Accelerator
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A Python chat web app that calls Azure OpenAI with managed identity, provisioned and deployed to Azure Container Apps via azd.
An Azure-hosted RAG accelerator that grounds a chat assistant in your own documents, returning answers with inline citations to sources.
An Azure-based reference accelerator where multiple specialized AI agents plan and execute user-specified organizational tasks, built on Microsoft Agent Framework and Azure AI Foundry.
A 21-lesson Microsoft course teaching developers the fundamentals of building generative AI applications with Python and TypeScript.
Hands-on .NET course with five lessons and runnable samples for building generative AI apps, agents and RAG.