Shata Solutions
← Back to marketplace
Customer Support

Ticket Classification

Auto-tag and route Zendesk tickets by topic, product, and urgency using NLP classification.

Automation overview

What this automation does

This automation starts when new Zendesk ticket created webhook. It openai classify topic & urgency and zendesk route to group, then low confidence tickets queue to triage Slack for human correction.

Trigger
New Zendesk ticket created webhook
Main actions
OpenAI classify topic & urgency → Zendesk route to group
Final outcome
Low confidence tickets queue to triage Slack for human correction
Execution time
30–90 seconds
Apps used
3 integrations
AI-powered
No
Difficulty
Intermediate
Intermediate
Workflow diagram

End-to-end scenario flow

Read-only visualization of how data moves from trigger to final result.

Trigger
Zendesk
New ticket created
AI
OpenAI
Classify topic & urgency
Step 2
Zendesk
Route to group
Result
Slack
Triage low confidence
Step-by-step

How each step runs

  1. 1
    ZendeskTrigger
    New ticket created

    Zendesk ticket create webhook

  2. 2
    OpenAI
    Classify topic & urgency

    NLP classifier tags billing, technical, shipping, or account

  3. 3
    Zendesk
    Route to group

    Auto-tag and assign to finance, tier-2, or logistics group

  4. 4
    Slack
    Triage low confidence

    Low confidence tickets queue to triage Slack for human correction

Apps used

Connected applications

Zendesk

Sends

OpenAI

Sends

Slack

Score

Summary

Automation summary

Trigger
New Zendesk ticket created webhook
Inputs
New Zendesk ticket created webhook, Business Name, Notification Email, Timezone
Outputs
Low confidence tickets queue to triage Slack for human correction, OpenAI classify topic & urgency, Zendesk route to group
Business goal
Automate ticket classification end-to-end with accurate app routing.
Execution frequency
Real-time on event
Estimated runtime
30–90 seconds
Error handling
Failed steps retry up to 3 times with exponential backoff; persistent failures notify your configured channel and log to the run history.
Dependencies
Zendesk, OpenAI
Benefits

Why use this automation?

Save ~4 hours/week

Remove repetitive steps from Ticket Classification so your team focuses on high-value work.

Reduce manual work

Connected apps sync data automatically instead of copy-paste between dashboards.

Prevent human errors

Standardized logic runs the same way every time, with retries on failure.

Requirements

What you need to get started

Required apps

  • Zendesk
  • OpenAI

Required accounts

  • Active Zendesk workspace with API access
  • Active OpenAI workspace with API access
  • Shata Automation workspace (free tier supported)

Required API keys

  • OAuth tokens for each connected integration

Permissions

  • Read access on trigger sources (forms, sheets, webhooks, or CRM objects)
  • Write access on destination apps (email, chat, CRM, or storage)

Estimated setup time

  • 20 min including app authorization and test run
  • Difficulty: Intermediate

Optional apps

  • Slack (optional)

Automation runtime

Activate this template

PartialSome steps run in preview until their integrations go live.

Connect required apps via OAuth or API keys. Credentials are encrypted server-side. Dev test runs mock external actions when connections are missing.

Draft
0%
Activation readiness

Est. setup time: ~5 min

Required connections

  • zendesk(required)
    Previewzendesk: workflow actions run in preview
    Missing
  • OpenAI(required)
    Missing
  • Missing connection: zendesk
  • Missing connection: OpenAI
Manage all connections →View all installations →

Workflow summary

Trigger
New Zendesk ticket created webhook
Steps
4 nodes
Required apps
2
Optional apps
1

Trigger

Webhook trigger

Activate this workflow to provision the trigger endpoint.

Run history

No runs yet. Install the template and run a test, or activate to receive webhook/schedule runs.

FreeIntermediatePartial20 min setup3.910.0k installs · 287 reviews
Customize for My Business
Analytics

Template performance

49.2k
Views
10.0k
Installs
310
Favorites
3.9
Avg rating
20.3%
Conversion
20 min
Setup time
93%
Success rate
87
Trending score
Most used integrations
ZendeskOpenAISlack

New Zendesk ticket webhook sends subject and body to fine-tuned classifier—billing, technical, shipping, account, feature request—with confidence score. Tags applied automatically; billing routes to finance group, technical to tier-2, shipping to logistics. Low confidence queues to triage Slack for human tag correction that feeds retraining export weekly.

Some steps run in preview until their integrations go live.

Required apps
ZendeskPreviewOpenAILive
Optional apps
SlackLive
All integrations
ZendeskOpenAISlack
Workflow preview

How Ticket Classification runs

1
Trigger
Load classification model
2
Process
Rule engine
3
Action
zendesk
4
Notify
Alert & log
ZendeskOpenaiSlack
Setup steps

Get started in 5 steps

  1. 1
    Connect Zendesk

    Webhook on ticket create; API for tag and group assignment.

  2. 2
    Load classification model

    Start with GPT-4o zero-shot; optional fine-tune on 500 labeled historical tickets.

  3. 3
    Map tags to groups

    Spreadsheet maps tag → Zendesk group ID and default priority.

  4. 4
    Enable feedback loop

    Human corrections in Slack export weekly to improve prompts.

  5. 5
    Audit 100 tickets

    Compare auto-tags to human baseline; tune until 90%+ accuracy.

Variables

Configurable settings

VariableTypeDefaultRequired
Business NametextYour BusinessYes
Notification EmailemailYes
TimezoneselectYes
Slack Channeltext#generalNo
FAQ

Frequently asked questions

Multi-issue tickets?

Primary topic wins; secondary topics added as additional tags when confidence >0.6.

VIP override?

Requester VIP flag forces priority urgent regardless of topic.

Spam tickets?

Spam classifier branch auto-solves and tags spam before routing.

Freshdesk instead?

Swap Zendesk connector; same classification logic.

Get started

Customize Ticket Classification for your business

Our team can adapt this template to your exact workflow, apps, and brand requirements.