Expanding Globally through a Serverless, Event-Driven Architecture on AWS with Playvox

Playvox provides workforce engagement solutions for digital-first contact centers, which are infused with artificial intelligence (AI). As a remote-first company, employees are based throughout the globe with core engineering hubs in Australia and Colombia. The company has been using AWS services since 2012; its original architecture was container based and cloud native, built on AWS.

“In those early days of being a startup, you want to choose a cloud provider that your customers are familiar with,” says Cheyne Wagner, senior vice president of global engineering at Playvox. “If a customer asks what we run on, and we say AWS, it gives our customers the confidence and comfort that they need.”

In 2020, Playvox set its sights on creating serverless, event-driven architecture to support its journey. A variety of business drivers were behind this engineering strategy. The company was expanding rapidly, both through acquisitions and by entering new marketplaces, and its customer base was growing globally. Customer needs were changing rapidly, too, growing in volume and complexity. Playvox wanted to introduce new features quickly to meet those evolving needs.

“The problem we look to solve for the contact center is complexity and the need to deliver support across many channels,” says Alex Bullen, chief product officer at Playvox. “We use technology to create AI-driven solutions that understand the complexities of omnichannel, particularly digital channels that have been growing exponentially.”

The company opted for a serverless solution to meet these demands, using AWS Lambda, a serverless, event-driven compute service that can run code for virtually any kind of application without provisioning or managing servers.

“By going serverless, Playvox can automate its software environment so that developers do not need to spend time provisioning or managing servers,” says Wagner. “As a result, the engineering team for infrastructure can remain small and agile, and it can instead focus on delivering business value for customers rather than managing underlying systems and architecture.”

Playvox dedicated 12–18 months to modernizing its workloads using AWS Lambda. By mid-2023, 80 percent of its workloads were running on serverless architecture. The company incorporated more serverless, event-driven products into the architecture, such as AWS Step Functions, a visual workflow service to build distributed applications, automate processes, orchestrate microservices, and create data and machine learning pipelines, and Amazon Simple Queue Service (Amazon SQS), a fully managed message queuing service for microservices, distributed systems, and serverless applications.

“With this architecture, we can build near-real-time integration into different solutions, and then we can surface that information for our customers,” says Wagner. “We can use the technology to offer a real-time experience across our omnichannel solution.”

Playvox is already seeing various benefits from the solution. Developers, who earlier had to spend time managing infrastructure, can now focus on developing features and enhancing the product for their customers. The time and work saved by going serverless and having the functions and scaling managed are some of the best benefits Playvox has seen so far.

Using AWS, Playvox has increased its scalability and speed while reducing costs. The company scaled to meet a 300 percent increase in demand over 1 year without needing to change any architecture. Speed improvements come from Playvox using AWS Lambda with Amazon Kinesis, which can ingest, buffer, and process streaming data in near real time. The company’s invocation cost decreased by three to five times compared with the costs associated with its old architecture, which had siloed each customer and required more resources to run. The company used these benefits to scale rapidly and reclaim employee time that had previously been taken up by managing and maintaining the product infrastructure.

The event-driven aspect of AWS Lambda, which previously received events from integrated services like Amazon Connect, Zendesk, or Salesforce, comes with its own benefits as well. Playvox developers can now plug event-driven models directly into the architecture to add new features or experiments, meaning that teams can innovate directly without affecting or being dependent on other product teams. This speeds up deployment time for new features, reduces time to market, and builds resilience. It has also made Playvox more agile so that developers can use events from around the solution without needing extra integration. “Speed to market has been a driving force for us,” says Bullen, “and this architecture has supported us in that.”