The client’s primary objective was to automate and systematize manual, fragmented sales processes through an AI assistant. The assistant had to: find and match offers with user preferences; coordinate and schedule interviews with time-sensitive notifications; and generate concise summaries. An important requirement was maintaining user identity and context across the web, Telegram, WhatsApp, and LinkedIn

Additionally, the assistant had to continuously gather information about the user from the internet, enrich and update their profile through chat interactions, and dynamically refine the information throughout the dialogue. It was also required to collect insights on investments, potential deals and opportunities, as well as evaluate companies and their business potential

We delivered a scalable architecture of 10+ specialized agents under supervisory orchestration. The platform unifies identity/state across channels, automates deal sourcing and tracking, and drives interview confirmations via push notifications while maintaining durable context and summarization for seamless handoffs

The system can also analyze users, match and connect them with each other to facilitate deal completion, provide real-time recommendations on what events to attend and who to meet, and fully manage the end-to-end deal process, acting as a comprehensive personal assistant

The automation reduced manual workload in sourcing and scheduling by about 45-60%. Interview confirmations became 2 to 3 times faster thanks to cross-channel notifications. User response rates in Telegram and WhatsApp increased by 20-30% compared to email-only communication