Many companies are touting the latest IoT and smart device to make your life better. However, most of these devices aren’t really “smart” but rather they are just connected to the internet so you can control them remotely. For example, to turn off your security system or turn on a light in your house, a user has to pull out his phone, open the app, log in, navigate to the right part of the app and then make the selection. Often this requires more work than actually just completing the task yourself and can cause users to get frustrated. Yes, remotely controlling devices is a great start, but to make devices truly “smart” you need to incorporate machine learning into them.
So, what exactly is machine learning? In layman’s terms, machine learning let’s products learn from a user’s past actions and then customize a more personal experience for the user. Because machine learning let’s a device learn from past interactions, these devices get smarter the more you use them. So, what are some examples of how we could see machine learning incorporated into smart products in the future? Well let’s look at a few cases.
Consumers are becoming more comfortable with having smart devices in their homes. According to Gartner, over 60% of all IoT products are used by consumers. One of the biggest areas for increased consumer value involves smart systems that can monitor energy usage and reduce maintenance costs Let’s say your HVAC system is connected to the internet and you went on a last minute vacation to Florida and forgot to turn down your heat. Through motion sensors, your house would detect that no one has been home for 2 days when typically, people are home. Your HVAC would send you a real-time alert that your heat is on. You could then remotely turn the heat down to minimize energy costs or the HVAC could just automatically turn itself down. This situation could also be applied to lights or your security system to name a few.
Also, machine learning could help a smart appliance like a refrigerator predict when a part may break or when a part needs to be replaced (does anyone really remember the last time they changed the water, ac, etc filter exactly when it was supposed to be replaced?). Once certain conditions are met, the consumer would then be alerted in real-time to replace a part or even have the ability to manually or automatically re-order the part.
The auto industry has many opportunities to make cars truly smart in the near term. For example, my husband has a 2006 Jeep Grand Cherokee. I hate it when I accidently grab his keys because as soon as I turn on the car with them, the seat moves forward and tilts back leaving me in an uncomfortable position. “His” key is programed to reflect his preferences, however, I often think how great it would be if the car would just know it’s me and not him. This is possible with machine learning. A machine learning enabled smart car could detect it’s me and not my husband that’s in either seat from a fingerprint scan or an image to detect who is in each seat. Not only would it know my seat position preferences, which it learns from how I’ve set the car in the past not because I programmed it, but it would also know how I like the radio volume, the radio station I like, the temperature, the position of the ac vents, if I want the sunroof open, etc. Other possible machine learning uses could involve automatically engaging 4-wheel drive if an incline or snow is detected, re-routing navigation if lights in the area are out or an accident has occurred, alerting a user if the air pressure in a tire is slowly decreasing, and adjusting settings in real-time that could lead to the engine overheating to name a few.
Streaming and cable services know what you’re watching and when you’re watching it. However, what is being done with that information? Right now, just product recommendations, but what could come next? Well, these services could combine viewing data with other information like a customer’s purchases (from another connected device) and browsing history to “know” what commercials to display to a consumer. This would lead to the viewer getting more relevant ads and TV services getting higher revenue per commercial spot for these highly targeted ads.
At the end of the day you need to provide solutions that provide value to your customers. This value could save them time or money or simply create a more personalized experience. By incorporating machine learning into smart products, manufacturers can provide this value to its customers to gain a competitive edge that will lead to more sales, more revenue and more customer loyalty.
Like our blog?
Join our mailing list to find out when our latest and greatest content is published.