Inspiration

A couple of my family friends got scammed by scammers over phone calls. They handed over the OTPs and personal documents to them not knowing the end consequences and lost a lot of money. Being into cybersecurity, this affected me a lot. These people were well-educated and tech-savvy but couldn't save themselves from messing up. This pushed me to think about how to build something to notify people beforehand if they are in between a potential scam or conversation that seems fishy!

What it does

Bastion receives data from the Omi hardware device as a transcription of the user's conversation. After carefully cleaning and using advanced AI tools, Bastion accurately determines if the conversation includes any malicious tasks or activities such as scanning QR codes, asking for personal identification, money, transfer, etc., and notifies the user in the form of a notification via the Omi mobile app in realtime.

How we built it

A concurrent workflow runs in the background consisting of identifying the user transcripts in the current session and, using AI tools, the transcripts are analyzed for malicious activities, potential threats, or sensitive information from the user.

Challenges we ran into

Accurately determining the potential of a scam with exact classification and generating a possible step to avoid getting affected negatively

Accomplishments that we're proud of

Building an impactful product that will be useful, especially for the elderly and people who are less aware of technological advancements, and making sure everyone is safe

What we learned

I learned a lot about building workflows, automation, scripting, and working with LLMs

What's next for Bastion - Fraud Analysis

Bastion aims to further expand and create a security awareness platform with further tools expanded beyond the Omi platform over to the user's mobile phone. Further features also include setting up emergency contacts and informing them in case of potential threats.

Built With

Share this project:

Updates