pendiva

Pendiva is an AI-powered commercial intelligence platform designed to help small businesses transform WhatsApp conversations into actionable business decisions. We built Pendiva using two key Meta technologies: Meta Model API, powered by Muse Spark, and WhatsApp Cloud API. Meta Model API and Muse Spark power the intelligence behind Pendiva. We integrated Muse Spark 1.3 through the Responses API to analyze customer conversations, identify commercial opportunities, extract relevant information, generate suggested next actions, and assist business owners with questions about their customers. We implemented structured JSON outputs, schema validation, controlled context retrieval, and source-message verification. Rather than simply generating responses, Pendiva connects AI recommendations to the original conversation evidence, allowing business owners to understand why an action is suggested. We also developed a Decision Brief experience that brings together customer requests, missing information, supporting evidence, and recommended actions in one place. WhatsApp Cloud API provides the foundation for receiving business conversations. We implemented webhook signature verification, message deduplication, and local conversation storage. We successfully tested incoming sandbox messages, although the current integration remains experimental and does not provide complete WhatsApp history synchronization. Our application combines these technologies with a human-in-the-loop workflow: AI helps identify what matters, but the business owner reviews recommendations and decides what to do next. Pendiva demonstrates how Meta AI can move beyond traditional chatbots and become an intelligent operational layer for businesses that already depend on WhatsApp.