The Mass Technology Leadership Council (MassTLC) recently hosted the The Value of Things (VoT): MassTLC IoT Conference in Waltham, Massachusetts. There was surprisingly little disagreement between the various panelists on the future of the Internet of Things (IoT), but probably not in the way you might expect. Sure everyone was in agreement about the billions of devices and trillions of dollars in market growth; however, something else was afoot. Nearly all the panelists were in agreement that the free-for-all model of the consumer Internet and mobile markets would not repeat itself. With IoT, the sheer number of devices and their pervasiveness has already put governments and regulators around the world on notice. Here's the consensus: design your system to support an opt-in model with high security and concerns for privacy otherwise, within the next 5 years, you will be shut-out.

How does one develop a user model for managing so many devices and complex networks? I think the trick is to leverage the already familiar analogs. Keynote speaker, Sanjay Sarma, Director of Digital Learning, MIT; Founder, MIT's Auto-ID Center spoke about connected cars. Cars have hundreds of devices in them today and no one thinks about each individual one. Instead people interact with the car on a higher level. Sanjay proposed a user model where communication occurs with avatars or in this case he called it a "Carvatar." Interaction is just like any other Avatar online, be it real or robotic. The intelligent car is then able to answer questions like "How much fuel do you have?" or "What is the cabin temperature?" However, the Avatar needs to have more capabilities than just simple factual statements, builders of these Avatars need to give them agency to ensure they have the ability to organize their subordinate devices into a set of questions and commands that supports a useful conversation and interaction:

Owner: Car, what is the cabin temperature?
Car: 78 degrees
Owner: Car, Set temperature target to 68 degrees.
Car: Ok

Collapsing all the interaction from dozens of individual knobs and buttons to a single smart interface with an agent is great, but it isn't enough. Sanjay also referenced the pet peeve many of us have with the current smart devices in the world today: they are designed for local pairing over say Wi-Fi or Bluetooth, which is limiting. I don't pair my phone to my wife's car because it will wipe out the pairing already configured for her phone nor can I switch control of the music to my daughter's phone while I retain control over navigation. In order for these Agents to be truly useful and innovative they must be able to have multiple and varied relationships with other devices.

This leads to my own vision: a social network of things that I believe is a user model humans can again relate with. Think about your device Avatar in terms of a social network. The device should be configurable to accept relationships with some parameters, and to build its own circles or lists (but with privacy and security settings that actually work!). For example, a device should have a circle called "owners" but also a circle called "siblings" where any device whose "owner" is the same as my own I can consider a "sibling." Imagine a circle such as "devices near me." Such a circle is less intimate so you probably want to share a lot less with that group of device "acquaintances." You can imagine using a simple set of rules to build out a handful of these circles. The devices can then go beyond simple self-organizing meshes for networking and develop more complex relationships of varying duration. Back to our car example, perhaps a car will broadcast "I'm braking" to members of the "nearby circle", but only "owners" can interrogate the temperature and command temperature changes.

Finally, what do these user models mean for the business model? For a start, if data is all opt-in and regulations prevent the reselling of data, you should plan to demonstrate the value of your application and charge for that value directly to the beneficiaries. Are you able to reduce repairs and maintenance? Offer a lower support contract for the up-front purchase of the application or instrumented devices. Does your application make cars safer? Turn insurance companies into channels and split the savings with them. Offer peace of mind to parents and pet owners with a park geo-fence that charges a per use fee as kids or pets enter the space. The business models will be as endless as the applications waiting to be brought to market and MetraTech is looking forward to supporting you all in this endeavor.