How I built an open-source WhatsApp claims bot for insurers in the DRC

Description and use

RapidOS enables insurers in the DRC to intake claims via WhatsApp, automatically extracting, grounding, and structuring data into claim files and PDFs, while remaining an AGPL-3.0 open-source solution.

Problem

In Kinshasa, claim submissions rely on informal WhatsApp messages that lose photos and require manual data entry, causing delays and errors.

Outcome

RapidOS enables insurers in the DRC to capture claims directly from WhatsApp, reducing manual entry, ensuring data integrity through grounding, and providing structured claim files and PDFs, all while remaining open source under AGPL-3.0.

Workflow

• 1. Insurer registers WhatsApp Business number and AI key in RapidOS.
• 2. Customer sends photos and text via WhatsApp to insurer’s number.
• 3. RapidOS webhook acknowledges quickly and dispatches messages per conversation order.
• 4. AI extracts claim fields using coverage definitions and grounding against customer text.
• 5. Bot asks for missing fields, updates draft, and confirms completion with customer.
• 6. Upon confirmation, claim is created, photos attached, PDF generated, and dashboard updated.
• 7. Insurer staff can pause bot, reply manually, and update claim status, notifying customer via WhatsApp.

Evidence

RapidOS is a WhatsApp assistant plus a dashboard. The insurer connects its own WhatsApp Business number and its own AI key, the assistant collects the claim in the customer's language, and the claims team gets a complete, structured file with a PDF.

Source

The handler acknowledges immediately and hands the message to a dispatcher. Messages from the same conversation (company + sender) run one after another, in arrival order.

Source

Grounding: every extracted value must be supported by what the customer actually wrote. A policy number, plate or date that isn't in the customer's messages is dropped before anything is shown or saved.

Source