A customer calls. A voice answers, understands the question, and handles it. No menu, no queue. It sounds like a demo at a trade show. But it can really happen today.

At the same time, we see many disappointments. Not because the technology is bad, but because expectations are not aligned. Time for a realistic view.

What already works well today

  • Recognizing and forwarding questions. An AI agent understands why someone is calling and directs the call to the right place. This often surpasses a menu.
  • Handling simple queries. Opening hours, status of a report, rescheduling an appointment.
  • Collecting information. Recording name, customer number, and question before a human takes over.
  • Summarizing. A neat summary in the customer dossier after the conversation.
  • Helping outside office hours. The customer receives an answer even when no one is available.

Where it still becomes challenging

  • Complex or emotional conversations. A complaint, a bereavement notice, or a complicated situation requires a human touch.
  • Language and dialect. Speech recognition has improved, but accents and jargon remain challenging.
  • Knowledge that is not organized. An agent is only as good as the information it relies on. Disorganized knowledge leads to disorganized answers.
  • The wrong role. An agent built as an internal assistant does not automatically behave as a customer-friendly receptionist. The conversation requires its own design.

Four rules for a responsible start

  1. Start narrow. Choose one common question and perfect it.
  2. Ensure a quick exit. The customer must always be able to talk to a human.
  3. Be transparent. Inform the customer that they are talking to an AI agent.
  4. Arrange privacy in advance. Decide what is stored, where, and for how long. GDPR also applies to AI.

What Microsoft technology is behind it?

Multiple components come together within the Microsoft ecosystem. Teams Phone or Dynamics 365 Contact Center for the conversation. Copilot Studio or Microsoft Foundry for the brain of the agent. And your own knowledge base and systems for the answers. The art is in connecting them.

Humans and AI together

The best results are not seen in organizations that want to replace people. We see them in organizations that free up their people for meaningful conversations. The AI agent does the groundwork, the employee makes the difference.

With 365 Agent, we help organizations responsibly implement AI agents in their customer contact. Start small, learn, and then scale up.

Want to know which customer queries you can already handle with AI? Schedule a conversation and we’ll see together where you can start.

Frequently Asked Questions

Do customers notice they are talking to an AI?

Often yes, and that's fine. Be transparent about it. Customers appreciate getting quick assistance.

Can the agent forward to an employee?

Yes, and this should be well organized. Including a summary so the customer doesn't have to repeat their story.

Is this safe for sensitive data?

It depends on the setup. With the right choices regarding storage, access, and retention periods, it can be GDPR-compliant.