Support split across seven inboxes, and a bot that cannot hand over
Customer conversations arrive on WhatsApp, Telegram, Slack, email, Mattermost, Instagram and Facebook, each with its own inbox and its own history. Bolt-on chatbots answer what they recognise and dead-end the rest, so the handover to a person arrives late, without context, and lands on whoever is free rather than whoever can actually help.
- Conversations fragmented across seven channels with no shared inbox or history
- Assistants that cannot tell when they are out of depth, so they escalate late or not at all
- Handovers that drop context and make the customer repeat themselves
- Teams locked to a single model provider, with no way to bring their own keys or switch models
Aalee Assist: one assistant, seven channels, and escalation decided by intent
Aalee AI's flagship product, pairing an AI assistant with a human team behind one shared dashboard and inbox. I lead Aalee Assist and built most of it, including the LLM, tool-calling and provider-routing layer, the embeddable web widget and the agent dashboard.
Seven-channel shared inbox
WhatsApp, Telegram, Slack, email, Mattermost, Instagram and Facebook in one thread view
Intent-based tool calling
The assistant decides when to answer, when to hand over, and which agent should take the conversation
Bring your own key
15+ LLM providers and many models, so a team uses its own accounts and switches without a migration
Embeddable web widget
Drop-in widget for any site, with desktop builds through Tauri and a progressive web app
Agent dashboard and inboxes
Assignment, history and full context carried through every escalation
Multi-tenant access control
Single sign-on and dynamic domain resolution keeping each workspace separate
Aalee AI's top product. It gives a support team one place to work instead of seven, and an assistant that knows when to stop and fetch a person.