As is true of people, some smart devices are smarter than others. And as is true with people, greater “smarts” in devices usually cost more time and money to produce. Call it the price of education. When building your own smart products and services, how can you best plan for this cost? And how much intelligence is really needed for your application?
The Ladder of IoT Intelligence
Intelligence in smart devices can best be described as an escalation of capabilities. While there are thousands of products touting themselves as smart, many of them aren’t particularly brainy. As an example, it’s one thing for a smart lock to open a door, quite another to only open the door for specific visitors at appointed times. Not every application needs this kind of sophisticated intelligence. The decision as to how much is necessary is a foundation of the value equation, for both maker and customer.
Here's a simple guide to escalating levels of smart capabilities:
Level 1: Perception
The most basic level of intelligence in smart devices is the ability to sense a current condition. That condition might be a relatively steady state, such as home climate, or a changed state, such as door opened or closed. Without applying higher levels of intelligence from the rest of the system (or from a human), the device is not very smart.
Level 2: Reportage
The next level up is the device’s ability to send the perceived information somewhere. The recipient might be a hub, another smart device, or the cloud, where additional intelligence or actions can be applied. The info is more descriptive and can range from real time data (it is now 72 degrees in the den) to a historical change (the thermostat has been pumping 5% harder vs. last month).
Level 3: Portrayal
Today’s most capable smart systems depend on rich data in order to perform wanted functions. If we think of data as news, and news as the “5 W’s” of who, what, when, where and why, this level of intelligence provides the most detailed presentation of the “what.” Here’s where the rubber meets the road in product development. A broad, extensible lexicon of data descriptors at this stage provides both needed granularity and future-ready capabilities.
Level 4: Prescription
Once the system has been fed with valid portrayals of status, what, if anything, should be done about them? Should there be any reactive machine-to-machine changes? Should the system send an alert and wait for human interaction? These rule sets are mission critical and often costly to develop. With this in mind, it’s worth remembering that human interaction doesn’t necessarily make a system less “smart.” Oftentimes the reverse is true.
Level 5: Prediction
For many smart applications, predictive abilities are seen as a holy grail. While we tend to think of this level of “smart” as highly sophisticated, it’s important to remember that we’re early in the curve of machine-based intelligence. Algorithm development and testing can be lengthy and costly, and an algorithm’s assumptions don’t spell truth – only a possible truth. Does your product or service need to make predictions? Is there real business value, or is it “nice to have?”
Level 6: Autonomy
A true autonomous system makes use of all the lower levels of intelligence and, through machine learning, can get smarter and more self-sufficient over time. Naturally, these are the most sophisticated systems, and cost the most in terms of development time and associated componentry. It’s worth asking again -- would this capability genuinely add value to your product or service? Value that will absolutely justify the added costs?
How Smart is Smart Enough?
In developing smart products and services, it’s fair to say that the most important goal is the wanted customer experience. What are their expectations? Customers may not know protocols or algorithms, but they do know what constitutes the desired value for their purchase consideration. Any rise in your cost of development means a corresponding rise in the cost of end purchase or subscription. Simply put, are your best features and benefits worth it to them at the end of the day? These aren’t technology decisions, they’re business decisions. And that’s where“smart” really comes into play.
# # #
Managing Partner Avi Rosenthal is often known as "the IoT Whisperer." His unique industry experience and deep connections with manufacturers, dealers and retailers has helped both startups and established companies successfully leverage the IoT revolution.