Project: AgriVisor - An Agronomic Advisor

[chetan2309]

Project Name

AgriVisor - An Agronomic Advisor

Description

AgriVisor - is a smart Agronomic advisor for farmers/growers. With advances in Agritech and continuous stream of agri products it gets difficult for Agronomist to remain updated about all the products that could be made helpful to growers. Agrivisor is a friend for all i.e growers, sellers/sales reps and most importantly agronomist. By ingesting most important agri products technical specifications as a data source to rag apps, a lot of insights can be drawn without having to memorise every minute details thereby creating a winning proposition for everyone i.e sellers, agronomist, and growers.

I have used online and freely available data source from web which is a technical product specification guide for a company - TODO

There are three major important features that I have introduced:-

1. Modification of prompt flows to make RAG behave as an expert agronomist thereby getting/generating better answers.
2. Added an additional flow to upload infected crop images from user interface. These images are the ones which can be addressed using products defined in the technical product specifications. Mostly related to crops such as Corn, Soybean and many more.
3. Since apis related to chat/ask are also available, I have also shown integration with other Microsoft technology such as Teams App. Teams App could be considered as an important integration as the framework allow us to use many of its own features like Voice search etc

This app can primarily benefit three category of personas:-

1. Sellers/Sales representative of a company can be made more informed about products of a company. This will give them an Out of the box knowledge tool to leverage and build on their knowledge
2. Agronomists - can help answer questions coming from growers about specific things about a product/condition/disease that they may want to zero down and may not be aware of. Agronomist like Sellers can also use this tool to expand their knowledge base about the products and offerings
3. Growers - Since it has support for voice search and multi-lingual support along with image search and normal text search it can hugely benefit a lot of growers across different geographies.
4. These agents rag apps can be quickly integrated across agronomic websites thereby unlocking a lot of value that can be provided to multiple user scenarios.

Technology & Languages

• JavaScript
• Java
• .NET
• Python
• AI Studio
• AI Search
• PostgreSQL
• Cosmos DB
• Azure SQL

Project Repository URL

https://github.com/csharma2309/azure-search-openai-demo

Deployed Endpoint URL

https://app-backend-v3od7n6qiv4m2.azurewebsites.net/

Project Video

https://youtu.be/mZb8nHQ0cfw

Team Members

chetan2309, csharma2309

[chetan2309]

Link to the source code of Teams App I demo - https://github.com/chetan2309/rag-teams-bot-aoai

This work is heavily inspired by yet another MSFT blog that I came across on how to leverage existing accelerators/patterns to make use of in short span of time.

[multispark]

Hello @chetan2309, thank you for participating in RAG Hack!

The team is working hard to distribute badges. Please have each team member fill out this form:
aka.ms/raghack/badge-dist

Thank you!

[chetan2309]

@multispark - I didn't this check for a few days. This form seems closed now. Can you please assist?

Regards,
Chetan

[multispark]

I've reopened, please fill out the form ASAP!