When AI moves from answers to action
Delivery delays gave Riachuelo an opportunity to move from AI-assisted work to agentic service.
As the leading reason customers contact its eCommerce operation, late orders generated a high volume of requests. Most customers wanted to know why an order was delayed, when it would arrive or what should happen next.
The questions were relatively straightforward. Resolving them was not. An employee still had to provide an update, share tracking information or create a support ticket.
Riachuelo built a Genesys Cloud Agentic Virtual Agent for its WhatsApp journey and trained it using scripts from some of its strongest consultants.
AVA can provide tracking links, create Level 2 support tickets automatically, handle critical situations and connect customers with relevant self-service options. It does more than answer a question. It moves the journey forward.
Unlike a traditional chatbot that primarily responds, AVA can understand the customer’s need and act on it within a defined service process.
AVA now retains 84% of these interactions in self-service, compared with 30% for Riachuelo’s traditional chatbot.
“We practically never received positive feedback when customers were served by a chatbot,” said César Nicoletti, Senior Manager at Midway and Riachuelo. “Now, with the AI agent, our customer satisfaction has increased by 60%.”
AI is no longer only analyzing an interaction or recommending what a person should do next. It is handling a defined part of the customer journey, while consultants remain available for exceptions and moments that require judgment or a human touch.
That is the shift from a human workforce supported by AI to a human plus AI workforce designed to work together.