What We Learned from Winning DisruptAI

ANNOUNCEMENTS / 05.31.19 / Jeff J.

Caption: We are taking wells to the MAX!

When we signed up for Microsoft’s DisruptAI hack-a-thon, we weren’t thinking about creating a unicorn startup or winning a piece of the $10,000 in prizes. We were there to learn.

Over breakfast, we made friends with data scientists, fintech veterans, startup enthusiasts, and many more brilliant professionals and students alike. We were feeling very overwhelmed by the competition. Kudos to Havas and Microsoft for assembling such a brilliant group of developers! There were even professional consultants and coaches from Microsoft at the event who hosted workshops and taught us to use Azure to bring our idea to life. When we had questions about IoT Hub or CosmosDB, they were incredibly helpful. And when they checked-in on our progress, their feedback helped us make our example a real, enterprise-ready deployment. Sure, landing 3rd place among 400 developers and 75 teams felt amazing, but that was just the cherry on top of a deeper and more lasting educational experience. 

Our Idea

Currently, water well monitoring systems are clunky, cost thousands to install, and thousands more to maintain. They lack LTE-M enabled IoT sensors and they lack responsive AI-driven analytics. This industry was begging to be reinvented and disrupted by new technologies that are on the market today.

Over 2 days, we built an IoT enabled well-monitoring system that uses AI to perform analyses on the well, and BlackBerry AtHoc to broadcast emergency alerts when the well is flooding or contaminated. Our AI model was trained on sensor data from our device. Once the model was trained, we called a weather API to grab upcoming temp/pressure/humidity data and ran this through our model to predict the water level in the well, weeks or months in advance. Our IoT sensor was built to connect to Telus’ LTE-M network so that we could report data to Azure IoT hub from areas with traditionally poor connectivity (remote, rural, and underground). In the end, we built an enterprise-ready service that could be deployed for less than 200 dollars. In production, we simply need to ship the device to the well owner. These could be farmers, municipal governments, etc. The customer attaches the device to the outside of their well and then we go on Azure and click-click deploy, or as Gurjit likes to say, ‘boom-boom, go’.

For a high-level view of our architecture, check out the image below. 

As you can see, our deployment was compartmentalized into pieces that could be easily separated and integrated into existing well monitoring systems. This was very important to us, because if someone didn’t want our entire A to Z solution, they could still choose one or more of our features to add to their systems.

Congratulations to the other winners who transformed Photoshop layers into physical pieces of art (1st place) and built a security feature for fintech that uses AI to verify a customer’s voice during support calls (2nd place).

If you would like to get in touch with us about this use-case or anything else, please don’t hesitate to reach out to us at and

For all other BlackBerry technology questions and comments, check out the BlackBerry Forums.

Jeff J.

About Jeff J.

As a part of the Enterprise Solutions Team, I work to bring the latest BlackBerry software and security features to life on the Android platform.