> ## Documentation Index
> Fetch the complete documentation index at: https://docs.orbit.devotel.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Inbox settings map — every operator console in one place

> A map of every console under Inbox → Settings — routing, tags, macros, dispositions, digital queues, auto-close, agent scripts, ticket automation, conversation intelligence, AI deflection, reply approvals, SLA — with onboarding order and a symptom-to-console troubleshooting table.

# Inbox settings map — every operator console in one place

The Inbox docs section splits coverage by surface — coaching, reply approvals, SLA escalation policies, social mentions — because that is how the features shipped. Operators do not work that way. A tenant rolling out an inbox team asks one question: *where do I set what?* This page answers it, collecting the twelve consoles under **Inbox → Settings** into one coherent map, in the order you should touch them.

Every console below is a tenant-owned control — you set it for your workspace, nothing applies until you turn it on. Each maps 1:1 to a tile on the **Inbox → Settings** hub in the dashboard.

## Console map

| Console                                      | What it does                                                                                                                                                                              | Docs                                                                                      |
| -------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------- |
| **Routing**                                  | Auto-assign conversations with round-robin, skills-based, and keyword rules — dry-run a rule before it goes live                                                                          | Dedicated page coming                                                                     |
| **Tags**                                     | Define the labels, colors, and descriptions agents apply to conversations                                                                                                                 | Dedicated page coming                                                                     |
| **Macros**                                   | Multi-step reply playbooks agents trigger with a slash-shortcut                                                                                                                           | [Macros — canned responses](/guides/macros-canned-responses)                              |
| **Reply approvals**                          | Require a supervisor to approve an agent's outgoing replies before they send                                                                                                              | [Reply approvals](/inbox/reply-approvals)                                                 |
| **AI deflection budget**                     | Cap monthly AI-resolved tickets, set a soft alert, and track savings vs human-agent cost                                                                                                  | Dedicated page coming                                                                     |
| **Auto-close stale conversations**           | Automatically close conversations idle for a set number of days, so the open queue shows only threads that need attention                                                                 | Dedicated page coming                                                                     |
| **Ticket automation**                        | When a ticket is created, tagged, goes idle, or changes status or priority, run actions like reassign, retag, set a field, escalate, or notify — with a dry-run before the rule goes live | Dedicated page coming                                                                     |
| **Agent scripts**                            | Guided, branching scripts that walk agents through chat, email, and messaging flows step by step, routing on the customer's answer                                                        | Dedicated page coming                                                                     |
| **Dispositions**                             | Outcome label sets agents and the AI assign when closing conversations, plus a distribution view                                                                                          | Dedicated page coming                                                                     |
| **Digital queues**                           | Skill-based ACD for email, chat, and social — route inbound conversations to the longest-idle skilled agent, with an SLA and an overflow policy                                           | [SLA escalation policies](/inbox/sla-escalation-policies) covers the per-queue alert path |
| **Conversation intelligence**                | Operators the classifier runs over every text thread to extract typed signals — a refund request, an escalation risk, a product category — that you can report on                         | Dedicated page coming                                                                     |
| **SLA report** (lives at Inbox → SLA report) | Set per-channel first-response and resolution targets and review live breach state                                                                                                        | [Inbox SLA timers](/guides/inbox-sla-timers)                                              |

There is no wrong first console — every tile is independent and off by default — but some orders waste less effort than others.

## Suggested onboarding order

When you stand up an inbox team, touch the consoles in this order. Each step builds on the one above it:

1. **Tags** — define the vocabulary everything downstream filters and reports on. A routing rule that matches on a tag only works once the tag exists.
2. **Routing** — set auto-assignment so inbound conversations land on a person instead of a shared pile. Dry-run each rule on recent conversations before enabling it.
3. **Digital queues** — if you route email, chat, and social by skill, define the queues and their overflow policies here. Routing rules can then target queues instead of individual agents.
4. **Auto-close** — pick the idle threshold now, not after three months of abandoned threads bury the queue.
5. **Macros** — write the reply playbooks once every channel is live, based on the questions that actually come in.
6. **Reply approvals** — if you gate replies for juniors or regulated queues, configure who is gated after the team has macros to answer with.
7. **SLA report / escalation policies** — set first-response and resolution targets per channel or per queue, then hang breach alerts off them.
8. **AI deflection budget** — once you know your real volume, cap AI-resolved tickets per month with a soft alert.
9. **Dispositions** — define close-out outcome labels early, so analytics and conversation-intelligence signals have categories to land in from day one.
10. **Ticket automation** — automate reassignments, retags, escalations, and notifications once the manual flow is stable. Dry-run every rule before it goes live.
11. **Agent scripts** — build guided flows for the conversations that follow a decision tree. Train them on live tickets only after routing and macros settle.
12. **Conversation intelligence** — define the operator taxonomy last: the signals you want extracted are clearer once you have real closed conversations to read.

Steps 1–4 unblock a working queue. Steps 5–8 make it efficient. Steps 9–12 turn it into something you can measure and automate.

## Concepts behind the consoles

These pages explain the underlying models the settings consoles configure:

* [ACD queue model](/concepts/acd-queue-model) — how queues, skills, and overflow fit together; this is what **Digital queues** and **Routing** assign against.
* [Agent presence lifecycle](/concepts/agent-presence-lifecycle) — the presence states an agent moves through, which determines when **Routing** and **Digital queues** consider an agent available.
* [Conversations](/concepts/conversations) — the conversation data model that **Tags**, **Dispositions**, and **Auto-close** mutate.
* [Conversation merge troubleshooting](/troubleshooting/conversation-merge) — when two threads about the same contact collide, merge decides which console's tags and dispositions survive.

## Troubleshooting: symptom → console

| Symptom                                            | Owning console              | Where to look                                                                                                                           |
| -------------------------------------------------- | --------------------------- | --------------------------------------------------------------------------------------------------------------------------------------- |
| Inbound conversations pile up unassigned           | Routing                     | **Inbox → Settings → Routing** — check rule coverage and dry-run against a recent conversation                                          |
| Routing rule never fires even though enabled       | Tags                        | Confirm the conversation actually carries the tag the rule matches on                                                                   |
| Open queue fills with abandoned threads            | Auto-close                  | **Inbox → Settings → Auto-close** — enable and set the idle-day threshold                                                               |
| Agent's replies never reach the customer           | Reply approvals             | [Reply approvals](/inbox/reply-approvals) — check whether the agent is gated and the queue is unworked                                  |
| SLA breach alerts never fire                       | SLA report / digital queues | [Inbox SLA timers](/guides/inbox-sla-timers) — targets must exist before the report can breach them; per-queue alerting is on the queue |
| Macro will not expand in the composer              | Macros                      | [Macros — canned responses](/guides/macros-canned-responses) — check the slash-shortcut and channel scoping                             |
| Ticket automation rule does nothing                | Ticket automation           | Re-run the rule dry-run in the console — the test output shows the predicate the live event missed                                      |
| Classifier produces no signals on threads          | Conversation intelligence   | The taxonomy has no active operator until you create one in the console                                                                 |
| AI-resolved tickets exceed the monthly budget      | AI deflection budget        | The soft alert threshold warns before the cap trips — review where the volume is coming from                                            |
| Agent asks for a structured path on a complex chat | Agent scripts               | One console owns the global script library; unused scripts do not appear in the composer                                                |
| Two threads about the same contact diverge         | Conversation model          | [Conversation merge](/troubleshooting/conversation-merge) — tags and dispositions from the merged thread resolve per the merge rules    |

## See also

* [Inbox setup](/guides/inbox-setup) — channels, assignment, and the workspace these consoles configure.
* [Reply approvals](/inbox/reply-approvals) — the supervisor review gate, in depth.
* [Per-digital-queue SLA escalation policies](/inbox/sla-escalation-policies) — queue-level alerting.
* [Inbox coaching](/inbox/coaching) — live coaching that complements scripted flows.
* [Social mentions](/inbox/social-mentions) — the mention surface these settings also govern.
* [Inbox terminology FAQ](/guides/inbox-terminology-faq) — the vocabulary used across every console above.
