CSAT 85%: how Parque Arauco achieved 24/7 service with generative AI
EMA, the AI agent of Parque Arauco, serves shopping centers in real time. Here is how the implementation went, the challenges and the results.
A shopping center receives all kinds of questions at all hours: opening hours, stores, parking, events, lost and found, services. Multiply that by several shopping centers and you have a customer service challenge that can hardly be solved with people alone.
That was the starting point for EMA, the Parque Arauco AI agent built on Databot.
The challenge
- High volume and high variety of queries, different at each shopping center.
- Constantly changing information: new stores opening, seasonal events, special hours.
- An expectation of immediate response from visitors who often write while they are at the mall.
The solution: an agent with context
EMA is not a decision tree. It is a generative AI agent that:
- Knows each shopping center, thanks to a knowledge base that the Parque Arauco team can update without relying on developers.
- Understands the intent behind the question, even when it is written informally or incompletely.
- Hands off to a person when the case requires it, along with the conversation context.
Results
“With generative AI, we achieved 85% customer satisfaction and 24/7 responses across all touchpoints.”
— Felipe Misle, Head of Corporate CX, Parque Arauco
- 85% satisfaction (CSAT) in the conversations handled.
- 24/7 coverage across all touchpoints.
What we learned
- The knowledge base is the product. The quality of the answers depends directly on how up to date and clear the information is.
- The CX team should own the agent. Being able to adjust instructions and content without an IT project speeds up continuous improvement.
- Measure from day one. Per-conversation satisfaction makes it possible to quickly spot which topics need more context.
Does your operation face a similar challenge? Let us talk.
