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

# Customer Support Tickets

> woku evaluates every negative piece of feedback and, when it warrants attention, creates an AI support ticket and sends it to your customer support platform

<Info>
  **Available on the Corporate plan.** This feature is part of woku's enterprise capabilities. [Talk to our sales team](https://woku.app/pricing).
</Info>

The **Support** module connects the voice of the customer with your support
team. woku evaluates every negative response it receives and, if it detects a
case that warrants attention, builds a **complete support ticket** and sends
it to the customer support platform you have configured. Your team receives
the case already assembled: what happened, what the customer is asking for and
what action is suggested, without anyone having to read and transcribe the
original feedback.

## Which responses generate tickets

A response enters the ticket engine when it meets all of these conditions:

* It is classified as **negative** (improvement opportunity).
* The customer is **identified** and did not choose to respond anonymously.
* It is a **production** response. Sandbox responses do not generate tickets.
* There is at least one **enabled destination** in the Support module.

Responses from **wokus, NPS, CSAT and CES** with a classified comment generate
tickets. Forms do not generate tickets, but their responses are part of the
customer history that the AI uses as context.

## Routing by trackers

Before going through the AI, the response is routed to a destination based on
its **trackers**, the external attributes that accompany each piece of feedback
(for example country, store or line of business).

* Each destination defines in which cases it receives tickets through
  conditions on trackers, grouped with **And also** and **Or if**: the
  destination matches when some **Or** group has all of its rows **And** true.
* Each condition uses the operator **equals** (an exact value) or **is any
  value** (all values of the tracker, without typing them one by one). To
  listen to several complete trackers, add an **is any value** condition for
  each tracker, joined with **Or if**. There is a maximum of 20 conditions per
  destination.
* The ticket belongs to the **first enabled destination** whose conditions
  match the response's trackers.
* If no condition matches, **no ticket is created**. There is no default
  destination.

## AI triage

With the destination resolved, the AI decides whether the case warrants
escalation. For this it receives the **complete customer history** (their
responses by tool, how many were negative and how long they have been a
customer) and the **AI context** defined in the destination.

If it escalates, the AI builds the case: title, severity (**high**, **medium**
or **low**), summary, description, history milestones, what the customer is
asking for, suggested action, churn risk and language. All the content is
marked as **AI generated**.

<Note>
  The customer's identity (name, email, phone) is added to the ticket
  deterministically and never passes through the AI model.
</Note>

## Delivery to the destination

Delivery is asynchronous and with **automatic retries**: up to 5 attempts with
increasing wait time between each one. The ticket timeline records **Sent to
the support team** only when the destination confirms receipt. In Zendesk, in
addition, the link to the remote ticket is saved, visible as **View in
customer support**. An email destination records the delivery to the team, but
does not save a remote link because the ticket travels as an email.

## Available destinations

### Zendesk

Requires an active Zendesk connection in **Company → Integrations**. When
creating the destination you choose the connected account and, optionally, a
Zendesk **Group ID** so that tickets reach a specific group. Check how to
connect your account in [Zendesk connector](/docs/en/integraciones/zendesk).

### Custom service

Sends the tickets over HTTP to any of your own systems. You configure:

* **URL** and **method** of the request.
* **Authentication**: none, username and password, Bearer token or API key in
  a header. The credential is stored encrypted and is not shown again.
* **JSON template** of the body, with `{{...}}` placeholders that are replaced
  with the case fields.

### Email

Sends the ticket to one or more email addresses, without depending on an
external integration. When creating the destination you add the list of
emails, up to 20 addresses. woku sends an email with the case summary,
priority, customer, comment, what the customer needs and a link to view the
ticket in woku. It does not require a connection or credentials.

## Configuring a destination

From **Support → New destination** you define the destination on a single
screen:

1. **Destination**: the type of platform and a descriptive name, for example
   `SAC Chile`.
2. **Connection**: the connection data according to the chosen type.
3. **AI context**: free text of up to 2000 characters that the AI uses when
   deciding and drafting. The clearer the context (what this destination
   covers, which cases are critical), the better the AI decides.
4. **Which tickets go to this destination**: the routing conditions by
   trackers, with a natural language preview of the rule you are building.

The conditions require trackers with data, which are managed in **Company →
Trackers**. Before saving you can use **Test connection** to validate the
credentials against the real destination.

Destinations can be **enabled and disabled** at any time. An inactive
destination does not receive new tickets.

## Tracking

The main **Support** view shows the operation by period (day, week, month,
year or all time): cards with the created tickets and their severity, and one
card per destination with its status. When you enter a destination you see its
detail and a table of tickets with name, tool, score, severity, AI summary and
date, with search and filters.

## Ticket detail

Each ticket shows the origin of the feedback, the score, the original comment
(with transcription if it was audio), the category, the summary, the case
built by the AI, the churn risk, the history and the customer data, and the
timeline with the delivery status to the destination.

You can edit the ticket's **title**, **severity**, **summary** and
**category** before or after delivery, for example to adjust the priority that
the AI assigned.

## More information

<Card title="Zendesk connector" href="/docs/en/integraciones/zendesk" horizontal>
  Connect your Zendesk account with an API token to feed the ticket
  destinations.
</Card>