Hi team,
We're building custom automations on top of BotSailor (via webhooks and the API), and we've run into a limitation that makes it hard to reliably distinguish who or what actually sent a message.
The issue:
Both the outgoing webhook and the Get Conversation endpoint return sender: "bot" for every outbound message, regardless of whether it came from the automated flow, an AI agent, or a human agent typing from the live chat panel. The agent_name field doesn't help here either — in our case it consistently returns the same static value for plain text messages no matter who sent them, while only structured messages (templates, interactive buttons, carousels) come back with agent_name: null. In practice, this means a plain text reply from our bot flow and a plain text reply typed manually by a support agent are indistinguishable in the data.
Why this matters:
Many of us are building human-handoff logic externally (pausing the bot when a human takes over, resuming it afterward, logging escalations, etc.), and that logic depends entirely on knowing the true source of each message. Right now there's no reliable field to build that on.
What we'd like to request:
For both the outgoing webhook payload and the Get Conversation API response, please add two fields per message:
- sender_type: indicating the actual origin of the message — for example
bot_flow,ai_agent, orhuman_agent. - message_type: indicating the format of the message — text, template, interactive/button, carousel, etc.
This would let developers build much more reliable automations on top of the platform without having to guess based on indirect signals.
Thanks for considering this — happy to provide more examples or test payloads if useful.