Air Canada GenAI Assistant
Leveraging Generative AI to create efficiencies for call centre agents
Air Canada agents handle a high volume of customer inquiries daily, often navigating through multiple systems and documents to find answers. IBM created GenAI ACpedia Assistant to quip call center agents with a guided flow interface to access the right information seamlessly, leading to faster resolutions and enhanced customer satisfaction.
MY ROLE
Lead designer
TEAM
1 Service Designer, 3 AI Engineers, 2 Developers, 1 PO
TIMELINE
3 weeks
Problem
Air Canada’s Contact Centre agents rely on ACpedia, a large knowledge base with over 2,500 documents, to assist customers. However, navigating this vast resource poses many challenges.
Time-Consuming Searches
Agents often waste time sifting through dense documents and data spread across multiple tabs.
Information Overload
Complex, text-heavy content results in inconsistent or incomplete answers to customer inquiries.
Task Overload
Multitasking while managing high call volumes (30–40 calls per day) adds to agent stress and impacts efficiency.
Opportunity
How might we create virtual assistant leveraging watsonx that analyze conversations, extracts intent, and generate easily consumable answers so that the agent can easily and quickly find answer for the customer, and ultimately reduce average handling time?
Outcome
After launching the ACpedia Assistant pilot, the tool was carefully tested with real Contact Centre Agents. It received overwhelmingly positive feedback for its ease of use, speed, and ability to improve efficiency. Agents noted a significant improvement in their ability to handle queries effectively, helping enhance both their productivity and customer satisfaction.
Faster Responses: Reduced call handling times to ~30% for short answers and ~40% for complex ones.
High Adoption: 100% of testers reported the tool as intuitive and effective.
Scalable Potential: Demonstrated readiness to expand to 2,500+ documents and multilingual support.
