SheriaBot

Welcome screen
SignUp
Login
Chatmain
Previouschats
Profile

Inspiration

The inspiration for SheriaBot came from a simple observation: the law, which is meant to protect and guide us, often seems inaccessible due to its complexity. I wanted to bridge this gap and make legal knowledge accessible to everyone, regardless of their background.

What it does

SheriaBot serves as a digital legal assistant, making the complex world of law and constitution accessible to everyone. Leveraging AI, SheriaBot transforms complex legal text into simple language that can be accessed by everyone, anytime, anywhere.

How I built it

I used Google Cloud’s Vertex AI Agent Builder to build a vertex Search App with no code. I created a datastore for the vertex search app that contains the PDF files of the Kenyan Laws. For this hackathon demo, I used two Acts only. I configured the app to search with follow-ups and use Gemini 1.5 Pro PREVIEW. I gave it instructions to customize the tone, style, and verbosity of the generative summary. I also enabled the extractive content. I integrated the search app to my frontend using API.

For the frontend, I used Flutter Flow to create a No Code App(chatbot) that can be easily deployed to app stores and web in one click. The chatbot makes the API calls to the Vertex Search App with the user query and receives back the response from the Vertex Search App as a summary generated by Gemini Pro 1.5 Preview.

Challenges I ran into

One of the main challenges was ensuring the accuracy of the information provided by SheriaBot. Legal information is sensitive and needs to be accurate. Ensuring this while also making the language simple was a tough balance to strike. Another challenge was I was not able to run this on the cloud maybe I needed a little bit more time with it to learn how I can run it on the cloud.

Despite these challenges, seeing SheriaBot help users understand their rights and navigate the legal world would be incredibly rewarding. I’m excited about the potential of SheriaBot to democratize legal knowledge and empower users.

Accomplishments that I am proud of

I used no code at all for this Project and I did it in three days only. The No Code approach made my workflows easy and fast and expanded my ability to build with generative AI. This demonstrates how easy it to build Generative AI apps like chatbots in a matter of hours using No Code or very minimal code.

What I learned

Throughout the project, I learned a lot about legal terminologies, the structure of laws and constitutions, and how to simplify complex information. I also deepened my understanding of AI, natural language processing, and chatbot development. I also learnt that the chatbot can used in any language as Gemini Pro 1.5 Preview does a good job of responding in the language it is queried in.