Zendesk AI Reply Draft and Tagging

Overview

This flow speeds up Zendesk support responses with AI kept under human control. On a schedule it reads open Zendesk tickets, sends each ticket's context to Azure OpenAI to draft a suggested reply plus suggested tags and a category, posts that draft as a private internal note on the ticket for the agent to review, applies the AI-suggested tags via the Zendesk connector, and notifies the support team in Microsoft Teams.

Why it matters: Agents spend significant time drafting routine replies. AI suggestions cut handle time while the agent always stays the decision-maker — nothing is ever auto-sent to the customer.

Built: CF-716 · Solution FlowLibsZendeskAIReplyDraftTagging · Flow ships Off (demo).

Use Case

A support team wants AI-drafted replies and automatic tagging as agent aids. A draft appears as an internal (non-public) note on each open ticket; the agent reviews, edits and sends it. Tags are applied automatically to improve routing and reporting.

Flow Architecture

Environment Variables

Connectors & Connections

Connector	API name	Actions used
Zendesk	shared_zendesk	GetItems PatchItem
Azure OpenAI	shared_azureopenai	ChatCompletions_Create_2024Feb15Preview
Microsoft Teams	shared_teams	PostMessageToConversation

Note — All connections are referenced as solution connection references; the flow is portable between environments as long as a connection is mapped at import time.

Customization Guide

Almost every realistic variant of this flow can be implemented by changing environment variable values. A few cases require small edits inside the flow definition — those are called out explicitly below.

Grounding / RAG

Point flowlibs_KbContextUrl at your Help Center, or extend the prompt to retrieve KB articles for stronger grounding.

Tone & brand voice

Edit varSystemPrompt (an Initialize Variable, no flow-logic change needed).

Tagging

Set flowlibs_AutoTag to false to suggest-only without writing tags. Tags are applied as a space-separated string via PatchItem; adjust to your Zendesk tag conventions on go-live.

Cadence / scope

Change the Recurrence interval and flowlibs_ZendeskTicketFilter (e.g. only status:new, or a specific group/brand).

Event-driven

Replace Recurrence + GetItems with the Zendesk GetOnUpdatedItemsV2 trigger for near-real-time drafting.

Key Expressions

The flow is intentionally light on Power Fx / WDL gymnastics — the heaviest expressions are the branch-name concatenation and the approval outcome check. They are listed below in the order they appear in the flow.

@first(outputs('Draft_Reply_With_AI')?['body']?['choices'])?['message']?['content']

body('Parse_AI_Response')?['reply' | 'tags' | 'category']

@{concat('https://', variables('varZendeskSubdomain'), '.zendesk.com/api/v2/tickets/', string(items('Apply_to_each_Ticket')?['id']), '.json')}

@{join(body('Parse_AI_Response')?['tags'], ' ')}

@equals(toLower(variables('varAutoTag')), 'true')

Customize & download