A few years ago, the phrase “smart home” might have conjured up the image of a few people (most likely Ph.D.’s) on leather couches reading philosophy or theoretical physics, with a backdrop of fully stocked bookshelves. Today, we know a smart home has nothing to do with intellectual pursuits – and maybe more to do with bookshelves that know what the homeowner likes to read.

So, what is a smart home, and what makes a home smart? Is it the numerous devices that mutually communicate via a network? Is it the autonomous systems, such as Alexa or Google Home, that analyze data to make pertinent decisions? Or is it a combination of both?

The answer is “none of these.” A home full of connected devices is not a smart home. A home is smart only if it makes living in it easier, safer and more fun. The “smart” element is the collective experience made possible by the coming together of various parts in the home of connected devices.

So if sensor-based devices, such as thermostats or lighting controllers or TVs, don’t behave as they should, then they don’t qualify as “smart.” For instance, the smoke detector in a smart home should set off the fire alarm as soon as it detects smoke, and even trigger an emergency call through a nearby smartphone while transmitting a video snippet of the situation to the emergency officer. If it doesn’t, the fallout could be serious – beyond the erosion of customer trust. So how can enterprises ensure their smart home solutions won’t fail them or their customers?

From ‘Working Features’ to ‘Fitness of Use’

Let’s take a moment to understand why traditional QA approaches don’t work here. There’s inherent complexity in an ecosystem of connected devices, or what we know as the Internet of Things (IoT), where every device functions in conjunction with every other device. Ensuring the quality of one such device depends on all the other devices in the ecosystem – a task that’s way too convoluted for a traditional quality assurance (QA) approach.

The answer lies in a smarter QA strategy that goes beyond the technical or functional aspects of ensuring customer experience with respect to real-world scenarios. Such a strategy needs to take into account the entire process, including the stimulus, the environment, the call to action and the required end result. It needs to validate not only whether features work but also the wider and broader aspect of whether they’re fit for use.

From Working Hard to Working Smart

Our work with a leading energy and utilities company in the UK illustrates this strategy for ensuring a smart home solution. Our approach incorporates the following mandates:

  • Devices should talk: The basic feature of an IoT-based ecosystem is its connectedness – several smart devices mutually communicating to get other devices in the ecosystem operating in a certain situation, or what we call interoperability. For us, the first step toward ensuring a smart home is validating the interoperability of a given environment.
  • Devices should respond: Smart devices in a smart home need a stimulus to respond as they should. For example, a fire alarm should go off only when there’s a fire. Ensuring the quality of a fire alarm wouldn’t be possible without a fire situation. Would that mean testers now have to risk their lives?
  • Devices should act independently: Not quite. This is where QA for smart homes leaps traditional boundaries to become smarter. The stimulus that needs to come from a situation or a human action can be substituted by smart test machines that eliminate the need for human intervention.

We simulated the home automation solution in a test lab to ensure the right behavior and the interoperability of the connected system. We also eliminated human intervention with our automated testing robot TEBOT, which mimics human-to-device interaction. The result: a fully automated, pay-as-you-go QA service, resulting in 50% faster releases, and the client’s smart home app being ranked number one in the iOS Store in the region.

The Extra Mile

Endeavors to make QA smarter go a step beyond automation and simulation. To be truly adept at handling the validation requirements of complex, interconnected ecosystems – and build in the right quality, the first time – QA needs to be collaborative, and it needs to be orchestrated so that tools and resources can be rapidly commissioned as needed to drive faster releases.

Karthikeyan Murugesan

Karthikeyan Murugesan

Karthikeyan Murugesan (Karthik) leads Cognizant’s Quality Engineering & Assurance Practice for the UK and Ireland. He consulted for more than 20 years... Read more