Yandle

Poster 1
Poster 2
Poster 3
Dashboard
Knowledge Base
Agent Configuration
Call History
Transactions
Campaings/Tasks
API
Org Details
Conversation Intelligence

Inspiration

In India, setting up a traditional commercial landline is a bureaucratic nightmare. Small business owners face mountains of paperwork, physical verifications, and weeks of waiting just to get a working phone number. During this time, they miss critical leads and lose revenue. We asked ourselves: Why can’t getting a business phone line be as fast as ordering food? This inspired us to build Yandle. We wanted to democratize enterprise-grade infrastructure for main street businesses, turning a multi-day telecom ordeal into a seamless, 5-minute digital onboarding experience.

What it does

Yandle is an AI-driven voice service for small businesses. In less than five minutes, any local shop can onboard and receive a dedicated digital landline. This line is instantly equipped with an intelligent AI Agent that handles incoming voice calls and chat queries 24/7, 365 days a week. It captures leads, answers customer FAQs, books appointments, and ensures that a business never misses a revenue opportunity again.

How we built it

We leveraged the full power of the Google Cloud Platform (GCP) to bridge the gap between traditional telephony and cutting-edge conversational AI:

Telephony Infrastructure: Integrated robust SIP trunks to provision instant virtual landlines for users upon registration, streaming audio directly into our cloud architecture.
Brain Engine — Gemini-powered intelligence: Uses Gemini models to understand customer intent, maintain conversation context, and intelligently route each interaction to the right response, action, or human agent.
Ultra-Low Latency via Gemini Live API: Utilized the Gemini Live API to achieve bidirectional, low-latency audio streaming. This enables natural barge-in support, allowing customers to interrupt the AI mid-sentence just like a human receptionist.
Hyper-Localized Indian Languages: Leveraged Google’s advanced audio and translation models to support regional dialects—including Hindi, Bengali, Tamil, Telugu, and Kannada—matching the caller’s language on the fly.
Backend Integration — AI-powered connectivity: Enabled Yandle’s AI agents to securely connect with business systems through APIs and webhooks, retrieve relevant information, and take actions such as checking availability, accessing customer data, and supporting automated workflows.

Challenges we ran into

Our biggest hurdle was regulatory and technical procurement. Securing a reliable SIP trunk provider in India while navigating strict telecom compliance and KYC frameworks took significant effort. It threatened to stall our early progress. However, overcoming this challenge became our biggest breakthrough. By successfully navigating this complex integration, we transformed a major launch roadblock into a powerful operational moat that protects our business from quick copycats.

Accomplishments that we're proud of

We are incredibly proud of shrinking a multi-day, paper-heavy telecom onboarding process down to just 5 minutes. We successfully achieved human-like, low-latency voice interactions with real-time barge-in support using the Gemini Live API. Additionally, building a system that can fluidly switch between multiple Indian languages ensures that technology serves the diverse reality of India's retail ecosystem.

What we learned

We learned that the true value of AI lies entirely in its accessibility. Building sophisticated voice models is meaningless if a small business owner cannot easily deploy them. By focusing heavily on reducing onboarding friction, we realized that solving unsexy infrastructure problems—like instant number provisioning and telecom compliance—is just as critical as optimizing the AI models themselves.

What's next for Yandle

We plan to scale Yandle into an omni-channel customer engagement hub. Our next immediate steps include integrating with WhatsApp Business and SMS follow-ups to send automated booking confirmations and payment links right after a call ends. We also plan to build deeper integrations with local Indian point-of-sale (POS) systems and inventory software, allowing the AI agent to check real-time product availability for callers.