Intent-based routing transforms customer support with AI in Dynamics 365 Contact Centre

Intent-based routing transforms customer support with AI (Leveraging Generative AI to automatically detect customer needs and route work items without manual rule-setting).

Real-time User Journey

This journey highlights how AI removes the “guesswork” from traditional keyword-based routing:

1. Customer Inquiry: A customer sends a complex, multi-sentence email: “I received my order #1234 but the screen is cracked and I’d like a refund or a replacement as soon as possible.”
2. AI Intent Detection: The Intent-based Routing engine analyzes the natural language of the email. Instead of just looking for the word “cracked,” it understands the underlying intent is “Product Damage” and “Returns.”
3. Automated Classification: The AI automatically tags the conversation with the correct “Intent” category and assigns a high-priority level because it involves a damaged shipment.
4. Precision Routing: The system identifies that this specific intent requires a “Returns Specialist” with “Hardware” skills. It bypasses general triage and sends the case directly to the specialist’s queue.
5. Agent Resolution: The specialist receives the case already categorized. They don’t need to spend time re-classifying the issue, allowing them to initiate the replacement process immediately.

Step-by-Step: How to Enable This Feature

Intent-based routing is part of the AI-enhanced Unified Routing suite. Follow these steps to configure it:

• Step 1: Admin Center Access

Sign in to the Customer Service admin center or Contact Center admin center.

• Step 2: Enable AI Routing

Navigate to Routing > Work classification. Ensure the toggle for AI-based routing is set to On.

• Step 3: Create Intent Categories

Navigate to Intents under the routing menu. You can either manually create intent categories (e.g., Billing, Technical Support) or allow the AI to suggest categories based on your historical data.

• Step 4: Train the Intent Model

Provide sample phrases or point the system to your existing case data so the AI learns how customers describe specific problems. (Microsoft provides pre-trained models for common service scenarios to speed this up).

• Step 5: Set Up Intent-based Rules

In your Workstream settings, go to Route to queues. Create a rule where the condition is “If Intent Equals [Product Damage]”, then route to the [Returns Queue].

• Step 6: Monitor and Refine

Use the Routing Diagnostics tool to see how the AI is classifying intents. You can “Thumbs Up” or “Thumbs Down” classifications to help the model learn and improve its accuracy over time.

Infographic: Keyword vs. Intent-based Routing

Capability	Keyword-Based (Legacy)	Intent-Based (AI)
Logic	Matches specific words (e.g., “Broken”).	Understands context and meaning.
Maintenance	Requires manual update of thousands of rules.	AI learns and adapts to new phrases.
Accuracy	High risk of “False Positives.”	High precision even with complex language.
Triage	Often requires human manual re-routing.	Fully autonomous classification and routing.
Setup Time	Months of rule-building.	Days of AI training/Pre-built models.

References

Intent-based routing transforms customer support with AI