Vision

<aside> 💡

We aim to automate 100% of debt-collection operations using AI

</aside>

Modern banks who lend to underserved segments have more bad loans because they are unable to follow-up and collect money from their large and risky customer base. Today, collections operations are manual, slow and expensive.

Our AI agents automate this entire process. We take loan books from banks and automate all the steps in the process - segmenting customers by risk, sending personalized AI followups across SMS, Calls and WhatsApp, negotiating chargeoffs, filing litigations and allocating agents for field collections.

Traction

Initially, we built an analytics engine using AI, trained on 150k+ calls, that can listen to a 2min collection call and predict customer risk with 85% accuracy. This is used by collection teams to strategize collections as per user cohorts., monitor agent performance and compliance in collection calls.

We have 1 paying customers - Nira Loans. We used our AI agents to improve their debt-collection rates to be at a 6 month high.

Demos

Analysis Platform

Youtube Link

https://youtu.be/zKVHsx36d_s

AI agents for automating collection calls

Youtube Link

https://youtu.be/wwFsmWleYVE

Product

Roadmap

1. AI Collection Agents:
   1. We have deployed AI Voice bots that we trained, using analytics of over 150k+ call recording - personalized to various user-cohorts - to maximize loan-recovery.
   2. Making 100+ calls per min - across all languages at a fraction of the cost.
   3. Our WhatsApp LLM agent recently started sharing payment links
2. Analytics Platform: This is our central engine to analyze calls/chats to predict defaulter risk and orchestrate next best step for collections. This helps:
   1. Identify the right channel, messaging and frequency to contact defaulter automatically.
   2. Catch defaulters early and act quickly to collect money back.
   3. Predict which borrowers are reliable, and save cost by not allocating agents to visit them.

Funding Status