Cross-channel handoff
Cross-channel handoff is the agent's ability to continue a single conversation across multiple messaging surfaces. A 7am voice turn references last night's text. A WhatsApp thread continues from a Telegram chat. The user picks the surface; the agent stays continuous.
Why it matters
Most users do not pick one messaging app and stay there. SMS at work, WhatsApp with family, Telegram for a hobby community, voice on the way home, Alexa in the kitchen. An assistant that lives in one app loses the conversation every time the user changes surfaces; the conversation has to be re-explained because the agent on the other side has no memory of what was said before.
Cross-channel handoff means the same agent, with the same memory, on every surface the user uses. The morning's voice conversation about a flight booking continues in the afternoon's WhatsApp thread asking about the boarding pass. The user does not re-explain. The agent does not forget. Identity is continuous.
The hard part is not the channels themselves; it is the identity model underneath. The agent has to know that "the SMS from this number" and "the WhatsApp message from this account" and "the Alexa request from this device" are all the same user. The memory has to update from any surface and serve any surface. Most consumer AI products are still per-app; the cross-channel pattern is what distinguishes a multi-channel agent from a chat product.
How Ajeris handles this
Each user has a single agent identity that listens on six channels: SMS, WhatsApp, Telegram, voice, Alexa, and email. The agent's memory writes from any channel and recalls from any channel. A user who texts in the morning and calls in the afternoon talks to the same agent with the same context. The handoff is invisible because it does not happen at the conversation layer; it happens at the identity layer.
