SurveyMonkey Verbatim Theme Analysis

Overview

This solution turns SurveyMonkey open-text (verbatim) feedback into categorized, sentiment-scored insight, then trends it weekly. It is built as two coordinated cloud flows in one solution:

1. FlowLibs - SurveyMonkey Verbatim Theme Enrichment (event-driven) — fires on each new survey response, pulls the open-text answers, classifies theme + sentiment + key phrases with Azure OpenAI, and stores the enriched record in a Dataverse table. 2. FlowLibs - SurveyMonkey Weekly Theme Digest (scheduled) — once a week, reads the last 7 days of enriched records, tallies mentions per theme, and posts a top-emerging-themes digest to Microsoft Teams.

Why it matters: Open-text is the richest feedback but the hardest to analyze at scale. Automated theming surfaces what customers actually care about — ranked, sentiment-scored, and trended — without anyone reading comments one at a time.

*As-built note:* classification is done with the Azure OpenAI connector; the description's mention of Azure AI Text Analytics is not part of the as-built connector set.

Use Case

A CX/product team collects verbatim feedback in SurveyMonkey and wants it themed, sentiment-scored, and trended automatically. Each response is enriched the moment it arrives; leadership gets a weekly Teams digest of the highest-volume themes. The two flows are decoupled through a shared Dataverse table (a correlationId is stamped on every enriched row so an individual response can be traced from enrichment into the digest window).

Flow Architecture

Environment Variables

Connectors & Connections

Connector	API name	Actions used
SurveyMonkey	shared_surveymonkey	OnNewResponseAddedSurvey GetResponseDetails
Azure OpenAI	shared_azureopenai	ChatCompletions_Create_2024Feb15Preview
Microsoft Dataverse	shared_commondataserviceforapps	CreateRecord ListRecords
Microsoft Teams

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.

Theme taxonomy

Edit flowlibs_ThemeSet to your own categories; the model is constrained to choose exactly one.

Auto-routing

Branch on flowlibs_theme (e.g. route Pricing to the billing team) by adding a Condition/Switch after the AI classification.

Trend / spike alerts

In Flow B, compare a theme's current count to a stored baseline and alert when volume spikes.

Sentiment escalation

In Flow A, add a Condition: if sentiment = negative, post an immediate Teams/Outlook alert in addition to storing the row.

Multilingual

Translate the verbatim (Microsoft Translator) before the Azure OpenAI call.

Digest cadence

Change the Recurrence frequency/interval, or filter the digest to negative-sentiment rows only.

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.

Choose exactly ONE theme from this allowed set: {flowlibs_ThemeSet}. sentiment must be exactly one of: positive, neutral, negative. keyPhrases is a comma-separated string of up to 5 short noun phrases.

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

flowlibs_responsedate ge @{variables('varWindowStart')}

@union(variables('varThemes'), variables('varThemes'))

length(body('Filter_Theme_Matches'))

Customize & download