This partnership dispels the myth that the only companies using Amazon Connect are newer brands looking to shake up the status quo.
No industry has more variables impacting customer service than airlines. Weather issues, mechanical problems, staffing, and other factors can delay or cancel flights. Bags can also get lost, prices can change, and planes can be oversold. This high level of unpredictability creates tension and customer service issues. Additionally, outdated systems, disconnected workflows, and limited self-service options have contributed to long wait times, customer frustration, and high operational costs.
Recognizing these challenges, Air Canada, the country’s largest airline, has been deploying new technologies to improve customer service across its passenger and cargo divisions. Using the Amazon Connect cloud-based contact center platform, Air Canada has addressed existing customer service gaps to ensure that the airline’s systems meet practical, real-world demands. Amazon Web Services (AWS) recently hosted a session with Air Canada at re:Invent 2024, where Sebastian Cosgrove, director of global customer service at Air Canada, shared how Amazon Connect is helping the airline innovate customer service across its business.
The most common issue people experience with contact centers is a lack of continuity, especially when moving from self-service to speaking with an agent. For example, when a person is transferred from self-service to a live agent, details about their issue are not always shared, so they’re forced to repeat themselves. Moreover, supervisors and managers often have limited data access, which hinders their ability to improve service quality.
Anything less than stellar service will escalate the stress level of an already anxious passenger, so those disconnected moments only make situations worse. One of the biggest factors holding companies back from modernizing is cost, and Sheila Smith, the principal Amazon Connect specialist for AWS, spoke about this on the panel.
She stated, “The single biggest cost in the contact center is not the cost of technology or network services; it’s the cost of agent resources. If we can find a way to increase containment through self-service experiences and deflect those (self-service) interactions from even coming into the contact center, that will drive a huge cost out of the business. It’s going to impact customer service as well positively.”
Staffing remains the lion’s share of the costs of running a contact center. According to AWS data, 75 percent of contact center costs is agent staffing, which drives organizations toward more economical options like self-service when possible. Most people (70 percent) prefer self-service over speaking to agents. However, when an issue needs to be escalated to an agent, poor handoffs typically lead to frustration.
AI has brought in a high level of fear of agent job replacement, but the reality is that companies like Air Canada can’t modernize until they cut people’s costs. The shift from people-heavy processes to digital methods aligns with current market trends. Gen AI enhances self-service across communication channels by guiding agents through complex issues and workflows. By simplifying workflows and reducing the need to switch between systems, Amazon Connect allows agents to pay more attention to complex customer needs.
In 2022, Air Canada launched a comprehensive modernization initiative, beginning with its Aeroplan loyalty program and extending to reservations and specialty services, enabled by its transition to Amazon Connect. This move integrated self-service capabilities into a unified customer experience, introducing a new interactive voice response (IVR) system with automated call recording, transcription, and sentiment analysis. Agents gained real-time transcription tools and the ability to flag issues easily.
For instance, Air Canada revamped the IVR system in its passenger division to improve the customer experience. When customers call, the system uses loyalty numbers to identify them. It then provides booking details and offers self-service options for easy tasks like selecting seats or checking baggage. For more complex issues, the system gathers context before connecting the customer to an agent, lowering the average handle time.
These changes have led to a 15 percent drop in call volumes, an 8 percent decrease in abandonment rates, and $4.7 million CAD in savings by reducing full-time positions. The updated IVR system has also handled 78 percent of informational queries without requiring agent involvement. According to Cosgrove, the shift made a noticeable difference in Air Canada’s customer service efficiency, allowing agents to focus on higher-value tasks.
“In November 2022, we were shy of 30,000 phone calls. Now, we’re consistently handling over 140,000 phone calls within the IVR. That’s the value Amazon Connect brought us and the benefits of self-service embedded in the IVR. We’ve started to realize these benefits,” said Cosgrove.
The airline’s cargo division faced similar challenges but required a slightly different approach. Air Canada replaced manual, email-based processes with Salesforce Service Cloud and integrated them with Amazon Connect. This created a centralized system that manages workflows automatically, giving agents a unified view of customer details and case history. The airline later added telephony with Service Cloud Voice, reducing average handle times by two minutes per phone call and increasing agent productivity by over 20 percent.
“Part of the reason we had such a huge gain was because before our integration with Service Cloud Voice, we had generic statuses such as away, ready, and training. Now we could mine the data to understand what the agents were doing and how we get them ready to always help our customers,” said Cosgrove.
Integrating Service Cloud has also significantly improved Air Canada’s quality assurance (QA) process. Previously paper-based, QA is fully automated, with results displayed on dashboards. This transformation has saved the airline 89 hours per month that were spent on manual QA processing in the past.
Air Canada is now focusing on the next phase of its customer service transformation. To help customers before issues are escalated to agents, virtual assistants powered by gen AI are being developed. These assistants will handle common queries and help streamline workflows. Guided workflows and issue summaries for live human agents are also in development, with feedback loops designed to refine the system based on real-world use.
Additionally, the airline plans to introduce AI-supported chat capabilities, further enhancing its omnichannel approach to customer interactions. These advancements aim to make the customer journey more seamless while equipping agents with the necessary tools. Air Canada is taking a careful and phased approach to modernization, which allows it to adjust without disrupting operations.
“When you’re talking about people, legacy systems, and multi-branch products, it takes time to make these changes,” said Cosgrove. “Anyone can quickly pull something out of the box, but if you want meaningful results, you must do it right first.”
While the case study at re:Invent was specific to the airline industry, the lessons learned can be applied to other verticals. By automating mundane, repetitive tasks, agents can spend more time and deliver more value on complicated tasks or with higher-value customers. This enables the brand to handle more interactions with fewer people—better service, lower costs, and happier customers—a win-win-win.
For AWS, showcasing a brand like Air Canada highlights how far the business unit has come. In the past, re:Invent’s Amazon Connect customers were brands like CapitalOne and Rocket Mortgage – the brands working to disrupt their industries. Air Canada is a blue-chip company looking to serve its customers better. At re:Invent, I met with banks, retailers, and other companies one would not put in the “disruptive” category. This dispels the myth that the only companies using Amazon Connect are newer brands looking to shake up the status quo.
Since the product launched, I’ve positioned Connect as a dark horse to watch, as I believe, given its part of the broader Amazon and has access to AI, cloud features, and more, it was on track to be a market leader. Last year, it moved into the Leaders quadrant on the Gartner MQ; this year, it rolled a bevy of mainstream brands. Amazon Connect is a dark horse no more.
