Problem Statement:
Imagine that you are going on a busy street and you see an animal in distress in the corner. Now, most of us want to help, but we could be apprehensive about an infection, bite or just don’t know who exactly to reach out or don’t want to get too involved. We are trying to address this by creating a direct, simple, and reliable bridge between people and the service providers who can take action - KarunAI. KarunAI is a combination of 2 words - Karuna + AI. In many Indian languages, Karuna means compassion. With the help of KarunAI, people will be able to perform an act of compassion which may go a long way.

KarunaAI is a multi-channel, AI-powered workflow app designed to bring urgent help faster whether it’s a suffering animal, a broken pipe, a public health hazard, or a vulnerable person in need. The main aim is to rapidly align each reported concern with the most capable responder, ensuring that every issue is received on time.
The platform combines open-source automation, powerful Voice AI, and advanced language models to create a smooth, human-friendly communication channel between the citizen, the AI agent, and the responding NGO or service center. The system ensures that the reporting process is as easy as sending a WhatsApp message, while the dispatch process is instant, automated, and professionally communicated via Voice AI.

The workflow is summarized below in the following steps:

1. A person sends an unstructured text (WhatsApp) describing the issue and location.
2. The Text AI Agent instantly picks up the message and extracts key details like: Type of issue (e.g., animal help, infrastructure problem, social welfare). Severity and specific details (e.g., injured dog, broken pipe, elderly fall).

Location information (address, or landmarks).

3. The information is then passed to the central workflow system for further action.
4. The n8n Workflow Engine receives the context and queries the internal service provider database (Google sheets in this case), retrieves the contact info of NGO/organization to identify the available NGO/Service Center.
5. The n8n engine triggers the Voice AI Agent to initiate an outbound phone call to the identified NGO’s emergency line. The Voice AI Agent uses ElevenLabs’ Text-to-Speech (TTS) engine to dynamically generate a natural, professional voice message containing the critical details: the precise location, the nature of the emergency, and the citizen’s contact details.
6. The n8n workflow sends an instant message back to the citizen (via WhatsApp/SMS) acknowledging the request, providing a Unique Tracking ID, and naming the NGO that has been dispatched.
7. Simultaneously, n8n sends a formal communication (e.g., Email) to the dispatched NGO containing the issue details, reporter contact details (If the reporter had agreed to being contacted).
8. The NGO physically arrives and resolves the issue on the ground.
9. An acknowledgement message is sent to the reporter citizen, thanking them and informing that help is on the way.
10. If the NGO/rescue organization is not able to send the help, then next organization conatct info record from DB/Google sheets is read and the n8n workflow makes call to the subsequent organiations.

Features:
High complexity involving multi-channel communication. It integrates a WhatsApp AI agent with a non-text communication service (voice call), requiring advanced orchestration (n8n) to manage data transformation, API calls, and conditional logic across the workflow.

The primary innovation lies in utilizing Voice AI & Chat AI of ElevenLabs to close the “last mile” communication gap. Instead of relying on slow, unverified emails or generic texts to the NGO, the system uses a professional, real-time voice call, guaranteeing immediate attention and clarity for urgent field dispatch.
The Proof of Resolution mechanism fundamentally shifts the model from a complaint system to an accountable civic intervention system, fostering public trust.

We also developed a basic UI for admin screen from where you can add new organizations to the DB/Google sheets. You can also look at metrics of how many total calls each org received and how many requests they handled and rejected.

Core Technology Stack of KarunaAI:
Workflow Orchestration - n8n (Node-based Workflow Automation)
WhatsApp AI Agent - Elevenlabs agents platform (Chat-only agent)
Voice AI Agent - Elevenlabs agents platform
LLM - Gemini 2.0 flash
TTS (Text-to-Speech) - ElevenLabs
Communication Gateway (Voice & WhatsApp) - Twilio API
UI - Bolt
Database/Storage - Google Sheets

Conclusion
The concept behind KarunAI is not limited to the social cause and reporting using text. It can also be applied to civic, industrial and enterprise use-cases. We can incorporate image processing/OCR, geolocation so that users can report issues using the image and share the location. E.g. if you see a dangerous infra issue, you can take a picture, share the location while reporting to the municipal co