Chatbots are one area of innovation that is driving and changing engagement for digital experiences, and a number of customers and others in the software market are adopting and embedding chatbots within their applications. Like other emerging technologies, chatbots add complexity to applications, which is why companies like Perfecto are trying to knock down that challenge and make it easy to deliver chatbot experiences faster.
Perfecto, a software quality lab, recently added a new set of capabilities that automate testing for voice-enabled chatbots and Facebook Messenger or Siri integrations. Its purpose is to provide developers building virtual assistant capabilities for functional correctness, responsiveness, and voice quality — shrinking the number of testing hours and giving them the resources the need to build voice-activated assistants.
The capabilities are fully integrated into Perfecto’s cloud-based platform, which includes all of the platforms for testing web, mobile and IoT devices.
“Testing software used to be a lot easier,” said Carlo Cadet, Perfecto’s chatbot market expert. “Now you are adding in this whole platforms’ conversation, like what device are you using versus what I’m using, what are the conditions you are using versus what I’m using, are you on Wi-Fi or mobile — as the engagement methods become richer, the complexity challenge rises for teams.”
The various tasks needed to accomplish automated testing for these services would require functions like the ability to convert a text string into audio, injecting audio onto the virtual assistant on the device, validate the response, and record and transact with the response.
(Read more: It’s a ‘Cognitive First’ world)
Mobile banking and insurance apps are using chatbots like Geico’s Kate, HSBC’s Andrew, and Bank of America’s Erica. These chatbots offer a core set of functions to customers, such as asking the questions like, “What is my balance?” Cadet said Perfecto wants to remove the manual testing component for development teams, and the way to do so is with automation.
For example, it is now easy to automate a sequence like, “What is my balance?” with Perfecto. Developers can code a string (“What is my balance?”), and flip that string to audio using text to speech, injecting the expected audio into the device. Audio responses from the in-app chatbot are converted to strings (speech to text) that can be compared to the expected response. In addition, developers can test for the audio quality (mean opinion score) that the chatbot delivered.
As brands enhance their applications with voice-based interactions, testing — specifically automated — becomes a critical component for creating and deploying chatbot services. According to Cadet, automation is the future, and companies need to have automation as the bedrock for continuous testing and innovating quickly.