I took an epic 10 minute voice call survey

On a random Friday morning while quietly sipping my tea, my phone rang at about 10am. The phone number was a landline unknown to me, so I hesitantly picked up the call thinking that it's perhaps my bank, a utility company calling me. It turns out, It was actually an automated phone survey.

At first I thought it was one of those telemarketing random calls trying to sign me up to their products, but when I had that I was about to take a 10 minuteHealth survey from Makerere University I believe school of health sciences, I reluctantly signed up.

For one, I was curious to know how people are using Interactive Voice Response(IVR) systems these days partly because I work with Africa's Talking. One of our products is actually voice. We provide developers with a simple and intuitive API to build voice-based applications such as this, call center and whatever else you can imagine.

The call survey lasted almost 10 minutes asking me particular questions about my health such with hypertension, alcohol and tobacco usage. If I were to rate how well done this survey was, I would probably give it a good 7/10.

The questions were precise and short albeit many. The recordings were crystal clear and they offered several language options. The options were concise; I was asked to select a particular number that corresponded to my answer of choice the same way you would with radio buttons on the web. When the caller needed a specific value from me such as number of days I ate fruits in the last week, I could easy enter the value using the phone's numeric keyboard. Here's a recording of the survey.

The caller was careful to structure his survey into sections and with the beginning or end of a section, they would tell remind me that I was about to or had finished answering questions from a particular section say Alcohol consumption habits. It was not any different from a normal web survey form with subheadings except this time I couldn't see what I was doing, but hear and only use my imagination to tell where I was in the survey.

Because of how professionally well thought out this survey was done, I finished it till the end to which I was entered into a raffle draw to win airtime worth Ugx 100,000. That could be a motivator for some people.

This call got me thinking again about what is becoming the next computer user interface. Six years ago when I wrote about how gaming-changing Siri, Apple's voice assistant was, there was no trace of Amazon Echo or Google Home and several other smart speakers. Now it every big Tech giant is rushing to build a voice interface to their mainly text-based mobile and web products.

Analysts believe that human-computer interactions are now going to be driven majorly by voice. You just speak to your phone or computer to get stuff done instead of clicking on a button or typing something with your keyboard. You simply have to say it.

If you want the lights turned off, simply say to Google Home "Ok Google, turn off the lights" or if you want to order bread, you simply tell Alexa "buy a loaf of bread". An order will automatically be placed from your Amazon account, money deducted from your credit card and bread delivered to your door step. Simple. That's the power of voice which is evidenced by over 40 million Alexa-powered devices sold already in 2017.

Voice-powered applications are gaining ground partly thanks to advances in Machine Learning, Artificial Intelligence(AI) and growth in computation power on portable devices as well as the cloud.

At the last Google I/O, Sundar Pichai, Google's CEO announced Google Duplex, a new technology for conducting natural conversations to carry out "real world" tasks over the phone. According to Google;

The technology is directed towards completing specific tasks, such as scheduling certain types of appointments. For such tasks, the system makes the conversational experience as natural as possible, allowing people to speak normally, like they would to another person, without having to adapt to a machine.

In the demo, an AI-powered automated system made a call to a actual person and made a reservation without them knowing that they were even talking to a robot. What? Yeah, that's how far we have come.

With the call survey by Makerere that I just took, it was obviously that I was interacting with a program running on a machine. I could not talk back my responses, instead I had to type them in a pre-defined manner. A slight mistake could either end the call or accept wrong answers.

But imagine an AI-powered IVR system similar to Google's Duplex?

I would have to respond to questions naturally like as if I was talking to an actual person on the other end of the line. The conversation would be less tense and more natural which would get me to give more meaningful responses than simply checking pre-determined lists. More people would be willing to take the survey and as a result Makerere would have a more meaningful dataset for their research.

In Africa where literacy levels particularly reading/writing are still low, it's obvious how significant voice apps powered by AI will be. Just this week Google announced that they will be opening the first-ever AI research Center in Africa based in Ghana. For us at Africa's Talking in the words of our CEO Samuel Gikandi, "we are hungry" about building the continent's digital fabric by empowering developers with intuitives tools and APIs to achieve that.

David Okwii

David Okwii is a Systems Engineer who currently works with Africa's Talking, a pan-African company serving millions of API queries for SMS/USSD/Voice, Airtime and Mobile Payments across 6 countries.

Kampala Uganda http://www.davidokwii.com

Subscribe to David Okwii dev blog

Get the latest posts delivered right to your inbox.

or subscribe via RSS with Feedly!