> For the complete documentation index, see [llms.txt](https://docs.11x.ai/llms.txt).
> Use it to discover all available pages before exploring further.

# Smart replies

Smart Replies categorizes prospect replies, drafts responses, escalates what needs a human, and automates follow-up. Test in simulation before enabling auto-reply.


Smart Replies handles what happens after a prospect responds. It **categorizes** the reply, **drafts** a response based on your guidelines, **escalates** anything needing a human, and **automates follow-up** so conversations keep moving. You set handling per category, and a **testing mode** lets you simulate conversations and review drafts before enabling auto-reply.

<Callout kind="tip">
Reply triage is where outbound actually stalls at volume. Plenty of tools help you *receive* replies; most still require a human to read every one. This is a throughput capability, not a convenience one.
</Callout>

## Reply categories

Categorization is based on patterns from thousands of campaigns.

| Category | Typical content | Sensible default handling |
| --- | --- | --- |
| **Information requests** | "Send me pricing", "What does it integrate with?" | Auto-reply from the knowledge base |
| **Product inquiries** | Specific capability questions | Auto-reply if covered; escalate if not |
| **Meeting requests** | "Happy to talk next week" | Auto-reply to book, always notify the rep |
| **Objections** | "We already use X", "No budget", "Not now" | Escalate early on; automate once drafts prove reliable |

## Handling options

For each category, choose one of three:

<Columns cols={3}>
  <Card title="Respond automatically" icon="zap">
    Alice answers directly. Use where your knowledge base reliably covers the question.
  </Card>
  <Card title="Skip" icon="minus-circle">
    No automated action. Use for categories you want handled entirely by humans.
  </Card>
  <Card title="Escalate" icon="user-check">
    Route to a person, with the conversation context attached.
  </Card>
</Columns>

<Callout kind="alert">
Escalation is only as good as its destination. Decide *who* receives escalations and confirm they see them — an escalation into an unwatched channel is worse than no escalation, because it looks handled. Wire it to [Slack](/integrations/slack).
</Callout>

## Testing before you automate

<Steps>
  <Step title="Run simulation mode" icon="flask-conical">
    Simulate conversations and review the drafted responses without anything sending.
  </Step>
  <Step title="Test the objections you actually hear" icon="message-square-warning">
    Not generic ones. Take your five most common real objections and check the drafts.
  </Step>
  <Step title="Test a question you know isn't covered" icon="help-circle">
    Confirm it escalates rather than being answered confidently and wrongly. This is the most important test on the page.
  </Step>
  <Step title="Enable per category, not all at once" icon="toggle-right">
    Start with information requests and meeting requests. Add objections last.
  </Step>
  <Step title="Keep sampling after go-live" icon="eye">
    Read a sample of auto-sent replies weekly for the first month.
  </Step>
</Steps>

<Callout kind="tip">
The failure mode to design against is a confident answer to a question your knowledge base doesn't cover. A reply that says "let me bring in a colleague" costs you nothing; a wrong answer about pricing or compliance costs you the deal.
</Callout>

## Guidelines and boundaries

Smart Replies drafts from the guidelines you provide, and from your [knowledge base](/alice/knowledge-base). The boundaries you set there apply directly here — claims Alice must never make, and topics that always route to a human.

Topics worth making unconditional escalations:

- Pricing beyond your published posture
- Contract, legal, or procurement terms
- Security and compliance specifics
- Roadmap or delivery commitments
- Anything from a named strategic account
- Anything from an existing customer

## What reps should do

Reps receive escalations and booked meetings. Their job here is feedback:

- **Flag a bad auto-reply.** That's a knowledge base or category-handling problem, and it's fixable.
- **Flag a missed escalation** — something answered automatically that should have come to them.

See [Getting started as a User](/help-center/get-started/getting-started-as-a-user).

## Frequently Asked Questions

<Expandable title="What should I do if an objection was answered badly?">
Either the objection isn't covered in the knowledge base, or that category shouldn't be automated yet. Add real objection handling — ideally from call transcripts — and move the category back to escalate while you validate.
</Expandable>

<Expandable title="Why is everything escalating?">
Categories are set to escalate, or your knowledge base is too thin to answer anything confidently. Check handling settings first, then knowledge base coverage.
</Expandable>

<Expandable title="Why is nothing escalating, including things that should?">
Handling is too permissive. Make the unconditional-escalation topics above explicit boundaries.
</Expandable>

<Expandable title="Why are escalations not reaching anyone?">
No destination configured, or it's going to a channel nobody watches. Connect [Slack](/integrations/slack) and tag the lead owner.
</Expandable>

<Expandable title="Why did the sequence keep running after a prospect replied?">
That's a sequence exit condition problem, not a Smart Replies one. See [sequences](/alice/sequences).
</Expandable>

<Expandable title="Why are meeting requests answered but not booked?">
Calendar isn't connected for the assigned rep. See [integrations](/integrations).
</Expandable>

## Next steps

<Columns cols={2}>
  <Card title="Sequences" icon="list-ordered" href="/alice/sequences" horizontal>
    Exit conditions when a reply arrives.
  </Card>
  <Card title="Knowledge base" icon="book-open" href="/alice/knowledge-base" horizontal>
    What Alice can answer, and what she can't.
  </Card>
  <Card title="Slack integration" icon="message-circle" href="/integrations/slack" horizontal>
    Where escalations should land.
  </Card>
  <Card title="Changelog" icon="clock" href="/changelog" horizontal>
    Smart Replies shipped in Episode 13.
  </Card>
</Columns>
