Why IoT and AI Projects are Doomed to Fail
Published 02/20/2018

Over the last several years, the hype over artificial intelligence (AI) and the Internet of Things (IoT) has exploded thanks in part to increasing computational power, smaller & cheaper sensors, and breakthrough advancements in related technologies. However, the fact of the matter is that most AI and/or IoT projects are doomed to fail, sometimes before they even begin. Even the projects that are accepted as "successfull" still miss out on the complete/bigger picture and it is there where Symbient has found opportunity.

Companies spends millions of dollars purchasing, developing, and training AI models that can detect, and sometimes even predict, issues much faster and more accurate than the best team of data scientists and analysts. Once an issue is detected or predicted, an alert or notification is sent in the form of an email, SMS text message, or simply a pretty chart on a user's screen.

Similarly, companies also spend millions of dollars purchasing and installing state-of-the-art sensor technologies in hopes of monitoring, analyzing, and interacting with their physical environment (or "habitat" as we like to call it). As these sensors identify issues within the habitat, alerts and notifications are again sent, sometimes in the form of a flashing light or sounding alarm.

However, one thing companies have in common is they all stop once the alert or notification has been sent. They then leave it to users, sometimes absent at the moment, to decide what to do to next to address the situation. You can think of this like the check engine light in your car turning on and leaving it to you, the driver, to decide whether to do anything about it or not. In the real world, the story doesn't stop when the check engine light turns on. In the real world, someone has to take action to correct the issue that initiated the alert. Otherwise, an already costly situation can become even more serious and in some cases life threatening.

At Symbient, it is our mission to take these AI and IoT projects to the next step by involving and augmenting human workers where appropriate. We do this by dispatching workers to the location where the issue and alert originated. Not only do we dispatch those workers, but we dispatch the right workers at the right time. As they are dispatched to perform work within their habitat, our technologies interact with those workers using speech recognition and synthesis technologies that allow workers to keep their hands and eyes focused on the task instead of fumbling with clipboards and/or mobile devices. This allows workers to perform more efficiently, safely, and reliably.

We do this using a thoughtful mix of Artificial and Augmented Intelligence, together with the Internet of Things, to bring and keep humans in the loop, especially after alerts and notifications are triggered by existing systems today. In many cases, keeping humans in the loop can mean nothing more than ensuring compliance by checking and guaranteeing all rules and regulations have been followed. In other cases, human involvement means performing maintenance before systems fail. Regardless of the reasons for human involvement, companies can longer consider AI and IoT projects successful by simply sending a notification and expecting someone, somewhere to figure out what to do next or to do nothing at all.

By ending the story when the check engine light turns on, companies aren't fulfilling their AI and IoT projects' potential, thus leaving them to fail. So, do something about it. Demand that further action be taken when issues are detected and predicted. Rely on your human workers more with their vast on-the-job experience and knowledge. And stop ignoring the check engine light - it can mean the difference between cruising the open road or sitting idly on the side while the rest of the world passes you by.