Multilingual Inbound Email Auto-Reply

Overview

This flow removes the language barrier from a shared inbox using Microsoft Translator V3 and Office 365 Outlook. When a new email arrives, the flow detects the sender's language, translates the message into the agent/pivot language (English) for an internal bilingual record, and replies to the sender with an acknowledgement localized back into their own language — all automatically, in 100+ languages. Why it matters: International inboxes either sit untriaged until a bilingual agent is free, or customers get a clumsy reply in the wrong language. Doing detect → translate-in → reply-out in one governed flow gives instant, on-brand acknowledgement while the agent works entirely in English. As-built note: The Translator connection is Microsoft Translator V3 (shared_microsofttranslatorv), whose Translate action requires a source language and has no Detect operation. Language detection is therefore done with a single built-in HTTP call to the Azure AI Translator /detect endpoint (a documented connector-gap fallback). All translation itself stays connector-first.

Use Case

Sales or support teams that receive email in many languages want an immediate, correctly-localized "we received your message" reply, plus an English rendering of the content routed to the agent queue for the real response. The flow ships Off; going live needs only connection authorization and environment-variable configuration.

Flow Architecture

Environment Variables

Connectors & Connections

Connector	API name	Actions used
Office 365 Outlook	shared_office365	OnNewEmailV3 ReplyToV3 SendEmailV2
Microsoft Translator	shared_microsofttranslatorv	Translate
HHTTP	http	POST /detect?api-version=3.0

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.

Skip internal mail

Extend the genuine-inbound guard to also bypass translation when the sender domain is internal.

HTML vs. plain

The body translation uses textType=html to preserve formatting; the acknowledgement uses plain. Adjust per channel.

Multiple agent queues

Route to different queues by detected language or sender domain by adding a Switch before Send Bilingual Record To Agent.

Persistent bilingual record

Add a Dataverse CreateRecord to log original/translation/detected-language for reporting (the current record is the agent email).

Trim quoted threads

Strip the quoted reply chain before translating to save characters and improve quality.

Faster polling

Lower the trigger recurrence below 3 minutes for near-real-time acknowledgement.

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.

@body('Detect_Email_Language')?[0]?['language']

@body('Translate_Email_Body_To_English')?[0]?['TranslatedText']

@body('Translate_Acknowledgement_To_Sender_Language')?[0]?['TranslatedText']

@and(not(equals(trim(coalesce(triggerOutputs()?['body/bodyPreview'],'')),'')), not(startsWith(coalesce(triggerOutputs()?['body/subject'],''), variables('varAutoReplyMarker'))))

@not(equals(toLower(coalesce(body('Detect_Email_Language')?[0]?['language'],variables('varPivotLanguage'))), toLower(variables('varPivotLanguage'))))

@{concat(variables('varTranslatorEndpoint'),'/detect?api-version=3.0')}

Customize & download