Deploy Azure Functions and Connect SitecoreAI Forms to Dataverse

In the previous article in this series, we built a production-ready Azure Functions backend for Microsoft Dataverse CRUD operations:

Azure Functions for Dataverse CRUD in .NET

In this fourth article, we complete the end-to-end integration by deploying those Functions to Azure, connecting a SitecoreAI Form through a secure webhook, creating an enquiry row in Dataverse, and retrieving enquiry records through a protected GET endpoint.

The final flow is straightforward:

SitecoreAI Form ➡ Azure Function ➡ Microsoft Dataverse

• Validate the Azure Functions project locally.
• Deploy the .NET isolated Function App to Azure.
• Configure Dataverse application settings securely.
• Make the create endpoint compatible with SitecoreAI webhook responses.
• Add a protected GET endpoint for enquiry retrieval.
• Create and activate a SitecoreAI enquiry form.
• Configure an API-key-protected SitecoreAI webhook.
• Test the complete flow and verify the Dataverse row.

The flow decouples frontend collection from your Dataverse data plane using a secure, serverless intermediary:

• The Azure Functions project from Session 3.
• An active Azure subscription with permissions to deploy Function Apps and Key Vault resources.
• A Microsoft Dataverse environment and the target enquiry table.
• An Entra ID App Registration / Application User with least-privilege Dataverse permissions.
• Visual Studio 2022 with the Azure development workload, or VS Code with Azure Functions Core Tools.
• Access to SitecoreAI Forms and Page Builder with permissions to manage webhooks.

📺 Video Walkthrough: End-to-End Integration

Before pushing changes to Azure, ensure the .NET 8/9 Isolated worker starts and executes the CRUD routes locally.

dotnet clean
dotnet restore
dotnet build --configuration Release
func start

Typical local endpoints are:

POST   http://localhost:7071/api/dataverse/enquiries
GET    http://localhost:7071/api/dataverse/enquiries
GET    http://localhost:7071/api/dataverse/enquiries/{id}
PATCH  http://localhost:7071/api/dataverse/enquiries/{id}
DELETE http://localhost:7071/api/dataverse/enquiries/{id}

Test the create endpoint:

curl -X POST "http://localhost:7071/api/dataverse/enquiries" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Amit Kumar",
    "email": "amit@example.com",
    "message": "Local SitecoreAI integration test",
    "test": false
  }'

curl "http://localhost:7071/api/dataverse/enquiries?pageSize=25"

1. Right-click the Azure Functions Project ➡ Select Publish.

2. Select Target: Azure ➡ Specific Target: Azure Function App (Windows or Linux).

3. Choose your existing Resource Group and hosting plan, or provision a new one.

4. Select the Subscription, Resource Group, Region, Storage Account, and Application Insights resource.

5. Click Publish and allow the remote build to complete.

Example Function App Base URL:

https://func-sitecoreai-dataverse-prod-uks.azurewebsites.net

local.settings.json is strictly for local execution. In Azure:

1. Navigate to Function App ➡ Settings ➡ Environment variables ➡ App settings.

2. Add your Dataverse application credentials:

SitecoreAI requires a predictable response contract. If the endpoint returns HTTP 200 without { "isSuccessful": true }, SitecoreAI flags a Negative response from the webhook error.

namespace SitecoreAI.Dataverse.Functions.Models;

public sealed record SitecoreWebhookResponse(
    [property: JsonPropertyName("isSuccessful")] bool IsSuccessful,
    [property: JsonPropertyName("message")] string Message,
    [property: JsonPropertyName("recordId")] Guid? RecordId = null,
    [property: JsonPropertyName("correlationId")] string? CorrelationId = null
);

Full Repository & Source Code

📱 Scan to access the complete SitecoreAI Dataverse Azure Functions repository
💡 If you want to review the full .NET Isolated implementation, request models, and function bindings, check out the complete project on GitHub:

Share your feedback or contribute to support the Sitecore developer community!

AmitKumar-AK/sitecoreai-dataverse-connector

Azure Functions-based integration connecting SitecoreAI (XM Cloud) forms to Microsoft Dataverse. Includes certificate-based authentication, CRUD operations, and console app demos.

C#

0

0

AmitKumar-AK/sitecore-ai-codebase

TypeScript

0

0

1. Navigate to SitecoreAI ➡ Design ➡ Forms.

2. Add input fields with matching keys:

   • txt_fullName (Single Line Text)

   • txt_Email (Email Address)

   • txtArea_enquiry (Multi-line Text)

   • list_feedbacktype (Dropdown / Choice)

   

1. Go to Form Settings ➡ Manage Webhooks ➡ Add Webhook.

2. Select API Key authentication.

3. Configure the following values:

   • Webhook URL: https://func-sitecoreai-dataverse-prod.azurewebsites.net/api/dataverse/enquiries

   • Header Name: x-functions-key

   • Header Value: [YOUR_FUNCTION_HOST_OR_FUNCTION_KEY]

4. Save the Webhook.

Click Test Webhook from the form settings dialog to verify payload serialization:

{
  "isSuccessful": true,
  "responseContent": "{\"Success\":true,\"Message\":\"Dataverse record created successfully.\",\"Data\":{\"id\":\"7634bbee-2996-f111-8075-70a8a5af01f4\"}}"
}

1. In SitecoreAI, open the target page within Page Builder.

2. Drag and drop the Form Component into the designated layout placeholder.

3. Select your newly configured Enquiry form from the properties panel.

4. Configure post-submit feedback (e.g., Thank you! Your enquiry has been received.).

5. Publish the page.

1. Open the page where SitecoreAI form added

2. Enter the required details and submit the form.

3. The form displays the configured success message and No negative webhook response is shown.

4. The payload contains the expected field keys.

5. Open Power Apps Studio .

6. Navigate to Tables ➡ Sitecore Enquiries.

7. Confirm that the new row exists with all columns mapped properly.

• Zero Direct Access: Dataverse is completely decoupled from the public frontend.

• Least Privilege: Entra Application User has access only to specific Dataverse tables.

• Separation of Concerns: Forms use POST-only keys; internal tools access GET endpoints via separate credentials.

• MCP Readiness: Clean input/output contracts enable straightforward exposure as Model Context Protocol (MCP) tools for enterprise AI agents.