And soon.
Over the past few years I have kept up with product trends and dialing in my patern recognition in the product world. Applying this knowledge to the ripples caused by artificial intelligence, it seems that our smart products are going to have to be even smarter — fast.
I’m sure you have interacted with a smart device recently. Fridges, light switches, and home assistants make up only a fraction of the hundreds of thousands of smart products we co-habit with. Usually, they are defined as a physical object (hardware) enabled by wireless enablement (software). This just happens to be my field of expertise, so lets dig in.
As previously mentioned, hardware and software are often put together to create a “smart product”. But with current trends, it looks like there is a new guest at the party.
When Apple, Google, and Amazon unveiled their personal assistants between April 2011 and November 2014, a tsunami of enabling products hit the market. These assistants became hubs for all peripheral devices allowing them to multiply in our environment.
Which brings us to today. We are familiar with human-to -computer interactions and have now gotten back to our average level of laziness — meaning we still want to do less. In the meantime, advanced generative AI models jumped into our house through the window, which we kind-of welcomed, and now we expect it everywhere.
But our smart products don’t seem to have welcomed it yet although the demand is present.
Which gets me to our current topic: In the next few years, smart products are going to naturally change definition.
So far, we needed to interact with products from a distance for them to be smart. In the future they will have to interact with us to be smart.
You and I have most likely run into this scenario. We are busy, need some assistance, call up to our assistant to enable another product like a fan, and since the prompt we gave it was not initially programed in, it failed. Now we try again this time raising our voice — again, no luck. You end up thinking to yourself, “why do I even have this thing if it is faster for me to press the button!”
As humans, we want to be understood. And when smart assistants don’t understand us it becomes very frustrating. A solution to these failed interaction is for our assistants to know us better, but this means we don’t just need AI, we also need a lot of ML. But both have a major drawback — privacy. For assistants to be better, they need data. To collect data, assistants need to hear us, even when we don’t want them to — and thats not cool. We are now at a crossroad.
So what is the solution?
There are a few. Excluding user education and Natural Language Understanding (NLU) as potential solutions since they involve deep machine learning and may run into privacy issues, we have to look elsewhere. The best way to improve our interaction with smart products is through Multi-Turn Dialogue and Feedback Mechanisms. By using these two techniques, our smart products could interact with us to better understand our command.
There is a lot more to be said about this topic and its solutions. It will ultimately come down to the investment companies are willing to make into better user-product interactions and which home assistant will be able to serve the end user best.
What this ultimately means is that smart products are going to need to become smarter. The best way to do this is to change the way they work. Which will ultimately change what they are, and thus their definition as well.
If you enjoyed this topic and want to stay in tune with the product innovation world, please checkout my written page for more articles.
Comments