Auto-Send Policy

Auto-send is how a reply reaches a customer with nobody reading it first. It is deliberately hard to trigger by accident: several independent conditions all have to agree, and any one of them can hold a reply back.

The rules#

A reply sends itself only when all of these are true:

  • The agent is on (not paused).
  • Auto-reply is on for that specific channel.
  • This conversation hasn't been paused or taken over by a human.
  • The reply's confidence is at or above the effective threshold.
  • No escalation trigger fired on the customer's message.

Fail any one and the reply becomes a draft in the inbox instead. Nothing is lost.

The threshold#

Every reply carries a confidence score from 0 to 1. The model proposes one; a gate then caps ungrounded and weakly retrieved factual answers so a leftover sales self-score cannot send. It does not raise a score into the auto-send band. Each agent has its own threshold, set on a slider in its Autopilot sidebar.

ValueMeaning
90%The default for a new agent - conservative on purpose.
70%The floor. Dragging below it asks you to confirm, because that's where replies the agent isn't confident about start going out unread.
Anything betweenYour call. Loosen it for a support agent answering documented questions. Keep it tight for anything that quotes a price.

Confidence is about grounding, not grammar

A fluent, confident-sounding reply built on no knowledge scores low - which is the point. If good answers keep scoring below your threshold, the fix is usually knowledge, not the slider.

Channel toggles#

Auto-reply is per channel, so you can let the agent answer website chat instantly while every email waits for a human. Website chat defaults to on for new agents - a live chat widget that takes an hour to answer is worse than no widget - and the rest default to off.

  1. 1Open the agent and find the Autopilot sidebar.
  2. 2Make sure the agent itself is on.
  3. 3Toggle auto-reply for Email, WhatsApp, Instagram or Voice individually.
  4. 4A channel with no connection bound can't be toggled at all - connect it first.

Pausing the agent clears every toggle

Turning the agent off doesn't just stop it - it switches off all channel auto-replies and puts voice actions back to draft. Turning the agent back on does not restore them. Set them again deliberately.

Per-conversation overrides#

Any single conversation can opt out of - or into - automation without touching the agent's settings. The conversation's own setting always wins over the agent's default, in both directions.

  • Taking over a conversation stops the agent replying there immediately, until you hand it back.
  • Auto follow-up can be paused for one thread while staying on everywhere else.
  • A conversation that has reached a terminal stage stops following up on its own, and says so.

Lead status changes the bar#

Lead statusEffect on auto-send
Active Client / ClosedAuto-send is blocked outright. Existing customers get a human, whatever the score says.
LostThe threshold is raised to at least 90%, so only a very well-grounded reply goes to someone who already said no.
Everyone elseYour configured threshold applies unchanged.

What overrides everything#

Escalation triggers are evaluated on the customer's message before confidence is even considered. If one fires, the agent hands over regardless of how confident it was:

  • The customer sounds negative or frustrated.
  • They asked for a human, in any of the ways people phrase that.
  • The message touches a sensitive topic - refunds, chargebacks, cancellation, legal threats, data deletion, fraud, or a request to stop being contacted.
  • You can add your own keywords to that list.

Why a draft was held#

Every held draft says why in plain language, so you never have to guess which rule caught it:

What you seeWhat happened
Confidence 62% - below the 90% auto-send threshold.It scored short.
Escalated to you for human review.A trigger fired.
Auto-send is off for this channel.The score was fine, the toggle wasn't on.
Paused - manual / agent off / stage.Automation is off for this conversation, and why.

Rolling it out safely#

  1. 1Run everything as drafts for a week and approve them by hand.
  2. 2Watch how often you edit before approving - draft acceptance in Analytics measures this.
  3. 3Turn on auto-send for your least risky channel first, usually website chat.
  4. 4Skim the Auto-sent filter daily for a fortnight.
  5. 5Only then extend it to email or WhatsApp, and leave the threshold at 90% until you have a reason to move it.

For the mechanics of how confidence is produced, see Auto-Send & Confidence.

Where the settings live

  • Agent on/off, channel toggles, threshold, voice action mode - the agent's Autopilot sidebar.
  • Take over, per-conversation automation - the inbox side panel.
  • Lead status - the lead record or the pipeline board.