Use AI in customer service
Add AI to customer service in layers: Copilot help for representatives, autonomous agents for case and knowledge work, and self-service agents over chat and voice that hand off to people with the full conversation.
The problem
Service teams juggle cases, chats and calls across channels, re-read long histories before answering and write the same notes and articles again and again. Customers wait in queues for answers that are often already in a knowledge base.
Self-service bots that cannot hand over cleanly frustrate customers, and fully autonomous handling of complex or sensitive cases is risky.1
The approach
- Assist representatives: Dynamics 365 Customer Service provides cases, knowledge base, unified routing, SLAs and the Copilot Service workspace, where Copilot summarizes cases and conversations and drafts responses.
- Contact center: Dynamics 365 Contact Center is a Copilot-first contact center that works with the CRM of your choice, across voice and digital channels, with IVR and AI agents, sentiment analysis and live transcription and translation.
- Autonomous agents: Microsoft lists a Case Management Agent, Customer Intent Agent, Customer Knowledge Management Agent and Quality Evaluation Agent; its 2026 release plan says these four reached general availability in October 2025. Autonomous agents hand off to a human representative with the full conversation history.
- Build your own: Copilot Studio's Customer Conversation harness uses a primary agent with specialized sub-agents and escalation paths to human representatives, published to website chat or a phone number and billed in Copilot Credits. Foundry Agent Service offers voice-based prompt agents through Voice Live, with telephony through Teams Phone extensibility or Twilio; Microsoft marks voice agents as preview.
When a simpler route is enough: if customers mostly ask questions answered in existing documents, a retrieval-based FAQ assistant is enough. Microsoft's agent guidance says static knowledge retrieval does not need an agent.1234567
Conceptual architecture
Diagram as a list
Copilot & business apps
- Customer channels: chat, voice, email, webWhere customers start conversationsConnects to Self-service agent (Copilot Studio or Contact Center)
- Representative workspace with CopilotSummaries, drafted replies and knowledge suggestionsConnects to Cases, SLAs and knowledge base (Dynamics 365 Customer Service)
Agents & automation
- Self-service agent (Copilot Studio or Contact Center)Answers routine requests and collects detailsConnects to Unified routing and intent detection
- Unified routing and intent detectionSends work to the right queue or agentConnects to Representative workspace with Copilot
- Autonomous agents (case management, knowledge, quality)Create and update cases, fill knowledge gaps, evaluate qualityConnects to Cases, SLAs and knowledge base (Dynamics 365 Customer Service)
AI models & platform
- Voice agents (Foundry Voice Live, preview)Real-time spoken conversationsConnects to Unified routing and intent detection
Data & analytics
- Cases, SLAs and knowledge base (Dynamics 365 Customer Service)System of record for service
- Knowledge retrievalGrounds answers in approved articles
Technologies and their roles
Microsoft Dynamics 3652
Service application and contact center
Customer Service and Contact Center provide cases, routing, Copilot features and autonomous service agents.
Microsoft Copilot Studio5
Custom customer-facing agents
The Customer Conversation harness runs multi-agent conversations over chat and voice with escalation to people.
Microsoft Foundry Agent Service6
Pro-code voice and chat agents
Voice-based prompt agents connect to phone channels without hosting voice orchestration code.
Microsoft Teams2
Expert collaboration
Representatives can collaborate with experts in Teams during live customer engagements.
What you need first1
- A clean, current knowledge base with owners for each article
- Defined escalation rules and the information a human needs at handoff
- Dynamics 365 or Copilot Studio licenses matching the features used (capabilities depend on purchased licenses)
- Channel setup: phone numbers, chat widgets and routing queues
- Quality measures agreed before launch, such as resolution and escalation rates
Risks and how to reduce them
- Agents give confident but wrong answers9
- Ground answers in approved knowledge, evaluate before launch and keep human escalation easy.
- Customers get stuck with a bot4
- Use agents that hand off with full conversation history to a human representative.
- Preview voice features in production6
- Microsoft marks Foundry voice agents as preview; check support and compliance before relying on them.
- Usage-based costs grow with volume5
- Track Copilot Credits consumption for Customer Conversation harness agents.
Documented examples
Air India
In productionAir India: AI.g customer service agent on Azure OpenAI
about 40,000 (2026), up from an average of 30,000 (2024) (measured)
Travel and hospitality, India
Related
Sources
- Welcome to Dynamics 365 Customer Service (opens in a new tab)
- Dynamics 365 Contact Center overview (opens in a new tab)
- Dynamics 365 Customer Service 2026 release wave 1 plan (opens in a new tab)
- Autonomous service agents in Dynamics 365 (opens in a new tab)
- Harnesses in Copilot Studio (opens in a new tab)
- What is Microsoft Foundry Agent Service? (opens in a new tab)
- Business plan for AI agents (Cloud Adoption Framework) (opens in a new tab)
- Overview of AI agents and Copilot features (opens in a new tab)
- Observability in generative AI (opens in a new tab)
Thank you. Your correction was sent.
The editors check it against the sources. If you left an email address, they may reply about it.