Ten years ago, the notion of on-demand services like fresh milk and food delivery, doctors making house calls, or hitching a ride in a stranger’s car would have been ludicrous—relics of decades past, unimaginable in our modern world. Even three decades ago they were a distant memory, eroded by a culture of process innovation relentlessly focused on delivering more—making it better, faster, safer, cheaper. Consumption of everything from milk to news had become part of a global economic engine, long before the Internet.
Since the Internet began gaining usage in the late 1990s, the holy grail of web entrepreneurs has been to take any common offline process, move it online, and watch usage grow. We could, it seemed, deliver services much like the “good old days” without sacrificing scale and efficiency. But as any number of spectacular flops from the dot-com boom demonstrate, the online services economy didn’t evolve as expected—if anything, the Internet’s maturation seemed to cement the increasingly location-agnostic nature of our existence. Despite the promise, the technology simply couldn’t deliver.
Today—finally—with the proliferation of smartphones and the application environment that came along with it, the online world is reshaping our offline world. The list of on-demand services emerging from enhanced modern technology covers almost every vertical: Uber for transportation, Instacart for grocery delivery, BetterDoctor for on-demand healthcare, and Handy for house cleaning, to name a few. Smartphones ushered in this Age of Service by turning humans into nodes of personal information, cues to location, and databases of physical behaviors. In so doing, they unlocked huge value in new service-based business models.
As we look beyond smartphones to the future of connected devices—smart products that do not require human direction but instead form a network of intelligent yet inanimate nodes—it’s critical that we remember the lessons of the past. To realize the promise of connected devices and unlock creative human services, there must be a platform upon which these services can be easily delivered. There are several key elements that I believe should be central to such a platform: intelligent image processing (computer vision), hyper-accurate location tracking, and low-powered sensors.
Intelligent Image Processing
Computer vision converts, processes, and understands images in a way that allows algorithms and decisions to be constructed around them. Not surprisingly, the computer vision engineer has rapidly become the most sought-after talent for connected devices startups. The ability to quickly process image or video data without large storage requirements or the need to capture minute detail (such as facial recognition, which is controversial given privacy concerns), would be an inherent component of a platform that delivers services. Companies like Placemeter, a computer vision technology that processes large amounts of video and turns it into structured data about the world around it, and Canary, another NYC-based startup, are perfecting computer vision technology. Companies like these could play an integral role in developing a common, basic set of features for a services platform.
Location tracking can be achieved via GPS, wi-fi pinging, and cell tower triangulation, but it is still not refined, especially in tight city blocks or in vertical spaces (such as multi-floor buildings). In addition, device locations are rarely networked together, which is a significant barrier to effective delivery of services. For example, it would be great if my home systems could know whether I’m outside or at my neighbor’s house and the moment I start heading home. Ideally a platform would stitch these various location technologies together and allow shared location information between my various devices. There are a few companies working on this, including a DC-based company called SocialRadar.
Finally, ultra-low-powered, programmable sensors with big chunks of memory are a must-have requirement for a platform that facilitates services being easily delivered. Psikick develops sensors that operate from 1/100th to 1/1000th of the power budget of other low power sensor platforms and they are equipped with energy efficient on-chip technology and are completely untethered to batteries. This means a low probability of losing data (or awareness) while a device is offline, and a high probability of receiving good service when a consumer is relying on service triggered by the device.
Just as it was impossible to predict exactly how the on-demand sharing economy would be unlocked, there remains plenty of uncertainty when it comes to artificially intelligent applications. Yet we can also look to the on-demand sharing economy as a useful model for identifying key challenges and core platform components in this new age of services—for starters, we know that these elements cannot simply be repurposed from a mobile platform. Today, developers have to rebuild these common requirements in each new device and in doing so, devote more time to the infrastructure than to the creativity around the differentiated service.
Without that creativity and differentiation, the era of connected devices could be relegated to a collection of “smart home” applications—temperature-controlled and perhaps well-secured with quick and simple video surveillance. That and a smart speaker system can’t be all there is—the Internet of Things is not about a few home appliances being the smart hub of communication and service. Connected devices and their potential are about a network of personal low-powered devices that will prompt consumer services—outside the home as much as inside the home—to be naturally and implicitly integrated into our daily needs.
NOTE: NEA is an investor in BetterDoctor, PlaceMeter, SocialRadar and PsiKick.