NoForma.AI

Inspiration

Most websites rely on contact forms to capture leads — but these forms often attract spam, go unanswered, or are abandoned by real visitors. As developers, we saw firsthand how hard it is for businesses to identify high-quality prospects from their own website traffic.

We built NoForma.AI to replace static forms with meaningful client interactions — using a conversational interface that helps businesses qualify real inquiries and reduce wasted follow-ups.

What it does

NoForma.AI is an AI-powered conversational interface that replaces traditional contact forms. Instead of typing into a blank form, visitors can speak to an AI agent through a video or chat interface. The agent captures key requirements, summarizes the interaction, and sends a structured summary to the company’s internal dashboard.

Companies can then review the conversation, assess client intent, and take appropriate next steps. Our goal is to help businesses focus on genuine leads — not spam or vague submissions.

How we built it

We designed NoForma.AI with a modular architecture and focused on delivering an end-to-end experience:

Backend microservices are built using Python Flask. These services handle communication with:
- Tavus (for video-based conversational interfaces)
- Sarvam.ai (used for summarizing client conversations)
- Cal.com (for calendar slot selection and meeting scheduling)
The company dashboard is built using Laravel with Inertia.js and React. This allows businesses to view incoming leads, access meeting summaries, and manage follow-ups from a central location.
A separate React application simulates the visitor-facing side of NoForma.AI — showing how a client would interact with the system from a business website.

Note: For this hackathon version, we did not implement persistent memory or a custom voice layer. We prioritized a clean workflow from client interaction to lead summary and business dashboard.

Challenges we ran into

Coordinating API integrations across multiple services (Tavus, Cal.com, Sarvam) under time constraints
Designing a clean dashboard UI that presents summaries, client info, and actions clearly and quickly
Structuring AI outputs into useful summaries without overengineering the prompt flow
Deciding which features to stub or mock for a functional demo

Accomplishments that we're proud of

Replacing the contact form with a working video-based conversational interface
Successfully summarizing client intent using Sarvam’s LLM
Building a functional company dashboard to manage leads and meeting data
Creating a live walkthrough that demonstrates the workflow from a client speaking to the agent → generating a structured summary → viewing it on the dashboard

What we learned

AI conversations are much more engaging than forms — but they need to be guided and focused to be useful
Properly formatting AI outputs into summaries that businesses can act on is as important as the conversation itself
Simple, well-scoped features are better than overbuilding within a limited timeline

What's next for NoForma.AI

Add memory support to retain context across multiple sessions with the same client
Offer a customizable widget that companies can embed into their own websites using plugins and SDKs
Add analytics and tracking to show conversion performance over time
Improve agent autonomy — allowing it to follow up, reschedule, cancel meetings or reroute leads with minimal manual intervention