LexiAI

Inspiration

It started one summer when my cousin, Tobi, asked me for help. He had just landed his first freelance gig—writing product descriptions for a startup. Excited, he signed the contract without a second thought. A few weeks in, things got weird. Payments were late, deadlines kept shifting, and the company added tasks he hadn’t agreed to. When he finally sent me the contract to look at, I saw the problem immediately. Buried deep in the fine print were clauses that gave the company the right to modify his scope of work without renegotiating payment. Another section locked him in for six months—whether he liked it or not. Tobi wasn’t careless. He just didn’t know what to look for. And truthfully? Neither did I. That’s when it hit me—why wasn’t there an easy way to scan a contract and understand what it really meant? That question turned into an obsession. And that obsession became LexiAI.

What it does

LexiAI is like a legal co-pilot for freelancers, startups, and small businesses. It:

• Scans contracts, agreements, and legal documents
• Extracts key details like payment terms, deadlines, and renewal clauses
• Flags risky clauses—so you know what could come back to bite you
• Summarizes everything in plain English
• read aloud extracted text
• translate to multi langauge

Tobi’s mistake wasn’t signing the contract. His mistake was signing it blind. With LexiAI, that never has to happen again.

How we built it

I knew AI could process documents, but I needed it to understand them.

• Azure Document Intelligence gave us the power to extract text and key details from contracts
• Azure Cosmos DB became our brain—a place to store legal insights and user data
• Azure Blob Storage handled document uploads securely
• Azure speech services handle text to speech, reading aloud contract clauses
• Azure translator handle transcibing to other languages
• Java Spring Boot tied everything together into a fast, scalable API

At first, the AI was laughably bad. It would pull random phrases and miss important details. But we trained it with real contracts, real case studies, real mistakes people had made—until it got smarter.

The moment of validation? When we ran Tobi’s old contract through LexiAI, and it flagged every single problem he had missed.

Challenges we ran into

Legal text is tricky. AI isn’t great at nuance, so we had to refine our extraction models. Contracts are never the same. Difficulty integrating azure text to speech live on deployment but was later solved using docker Security matters. Legal documents are sensitive, so we had to make sure data encryption and privacy controls were airtight.

There were times when I thought, maybe this is too big of a problem to solve. But every time we ran a contract through LexiAI and said, “we would’ve never caught that,” I knew we were onto something.

Accomplishments that we're proud of

• Helping freelancers avoid bad contracts
• Giving startups legal clarity without needing a lawyer
• Building an AI that understands real-world legal risks

What we learned

• AI can demystify legalese. The right tools turn complexity into clarity.
• Cloud integration is powerful. Azure’s AI, storage, and database services made LexiAI scalable and secure.
• The future of law is AI-assisted. Lawyers and businesses can work faster, smarter, and fairer with AI-powered insights.

What's next for LexiAI

This is just the beginning. Next, we’re adding:

• A chat-based legal assistant—ask questions, get instant answers
• A risk rating system—so you can tell at a glance how safe your contract is

I started this project because I saw my cousin get trapped in a bad contract. But I quickly realized he wasn’t the only one.

LexiAI isn’t just about reading contracts. It’s about giving people power over the fine print. And we’re just getting started.

Built With

• azure
• azureai
• azurespeech
• blob
• copilot
• cosmosdb
• documentcloud
• git
• java
• react
• springboot