The Opportunity
WOO, the telecommunications operator, wanted to improve customer service efficiency and reduce the number of phone calls, especially for repetitive, low-complexity requests. The goal was to offer a digital-first support experience that could scale with demand while keeping human agents focused on higher-value cases.
High call volume
Customer service teams were overloaded with calls, with a 55% handover rate to human agents. Many of these interactions followed predictable patterns that could be automated without compromising service quality.
Scattered information
Outdated and fragmented internal databases made it difficult to provide consistent and fast answers. Knowledge was spread across multiple systems, slowing down both agents and any attempt at automation.
Long response times
Clients often faced delays and inefficiencies, leading to slower resolutions and higher support costs. The existing setup made it hard to offer a truly digital, on-demand support experience.
The Solution
We built WOOGO, an AI-powered chat assistant for Woo’s website and app, designed to reduce call volume while keeping human oversight when needed. It helps customers and non-customers solve issues directly in chat, retrieving accurate answers from internal knowledge bases.
Developed side by side with internal teams, the project involved cleaning databases, updating FAQs, and building a monitoring and alert system that allows operators to step in whenever necessary. This not only ensured higher success rates but also greater trust in the tool.
The assistant distinguishes between generic and authenticated chats, providing secure access to client data when required and guaranteeing seamless handover to agents. Continuous retraining and feedback loops improve accuracy over time, reducing handover rates and making the assistant more efficient and reliable.
The Impact
With the AI chat assistant in place, WOO achieved a significant drop in handover rates from 55% to 29%, meaning 26% more conversations are now solved directly in chat.
Each month, the assistant handles around 56,000 conversations for more than 32,000 users, helping to contain about 40,000 requests that would otherwise reach call centers. This allows human agents to focus on cases where their expertise adds real value, while generating monthly savings of about €37k and over €500k since launch.
Customers benefit from faster service, with an average response time of 4.6 seconds and accuracy close to 90%. Beyond financial and operational gains, continuous retraining and feedback loops keep improving results, making the system more reliable and effective over time.
See it in action
