AI replies for inbound iMessage.
Auto-reply when it is safe. Copilot when a rep should approve.
When a lead writes back, Tuco can draft the answer for a human, route it to the owner in Telegram or email, or send it automatically in autopilot mode. It uses knowledge, thread history, HubSpot and CRM context, and AI SDR handoff rules so replies sound native instead of robotic.
Last updated:
Draft. Then approve.
Owners get the reply in Telegram, email, and Unibox before anything goes out.
Auto-reply when safe.
Low-risk inbound questions can be answered automatically using workspace rules, confidence thresholds, and KB context.
Keep the thread moving.
An AI SDR can take over when the reply should move into qualification, booking, or the next sales step instead of generic support copy.
Send it to the right rep.
Personal Telegram first. Workspace fallback second. Email backup always there.
Inbound thread
Multi-bubble reply with typing rhythm.
Hey, quick one. Can this sync replies back to HubSpot too?
Inbound reply
KB combines the last 60 seconds of inbound text before drafting. Owner found: Sarah.
Yep. Replies can log back to HubSpot automatically.
Bubble 1
You can also stop campaigns when a note, tag, or property changes.
Bubble 2
Inbound bursts are debounced first so the draft uses the full thought, not half the reply.
Replies split into short iMessage bubbles with typing indicators in between.
Choose the level of human approval that fits the risk and the volume.
Notifications target the contact owner before they fall back to the workspace channel.
Answers. Without the lag.
Replies follow contact ownership. Personal Telegram can receive the draft before the shared workspace channel ever sees it.
Manual for full control. Copilot for approve-or-edit. Autopilot for routine inbound questions that should move now.
Short multi-bubble responses with typing rhythm feel native to iMessage. Not robotic. Not over-written.
Guide a conversation toward qualification, booking, or close using stage objectives instead of one generic answer mode.
Voice notes, attachments, reactions, and thread history give the AI better context before it drafts the next move.
If a draft should not send, the owner still gets notified. If the model flakes, the thread does not disappear into a black hole.
Workflow
How replies move.
Inbound text comes in. Tuco does the boring part first.
Collect the inbound burst.
The system waits for the reply burst to finish so one answer can address the full message.
Load the right context.
Knowledge, thread history, owner, AI SDR context, and language rules all get pulled into the draft.
Draft or auto-send.
Copilot routes for approval. Autopilot sends right away when the workspace allows it.
Log the outcome.
Telegram, email, Unibox, HubSpot write-back, and activity history stay in sync after the reply lands.
Related
What it plugs into.
Keep the inbox moving.
The best AI reply engine is not the one with the longest answer. It is the one that gets the rep, the draft, and the next step right.