NetherlandsEnergy and utilities2024 – ongoing

Eneco

AI chat agent built in Copilot Studio

Eneco: new AI agent handles 67% of chats without hand-off, up from 40%

WorkingVendor's case studyVendor ↗

Problem

Call volumes kept rising and the energy retailer lacked agents to answer them. The old chatbot was hard to train, so simple questions still reached call centers.

Approach

Eneco built an agent in Microsoft Copilot Studio with Azure AI language understanding for intent recognition. Partner Seamly linked it to live chat. When unsure, the agent hands over to a person with an AI-written summary.

Results

67% vs 40%Conversations handled without live-agent hand-off, new agent versus previous chatbotVendor's case study[1]

24,000Chats per month handled by the AI agent, versus an average of 10,000 with the previous chatbotVendor's case study[1]

3 monthsTime to deploy the agent including live chat integrationVendor's case study[1]

Takeaway

The new agent handled more chats and a higher share without staff hand-off. Microsoft reports 67% against 40% for the old chatbot, but also gives a different figure of 70%.

Sources

Eneco customer story: Copilot Studio
Microsoft5 Mar 2025Vendor publication

Source summaries

[1]Microsoft5 Mar 2025

Eneco replaced a weak chatbot with a Copilot Studio agent in three months. Microsoft compares chat volume and hand-off rates with the old bot.

Full details are in the original sources.

Added 25 Oct 2026 · Updated 25 Oct 2026