Cognition in IoT — Built World

I have the pleasure of working with some very talented engineers, data scientists and developers in the software and analytics space.

The concept for our Cognitive-X, consulting business started with the end customer and worked back to the tech required to deliver.

As we come from B2B backgrounds, the business case and ROI is always at the forefront of our thinking. If there isn’t a business case, or a measurable return, we cannot get behind the offering, so as always, we use the business case as our primary benchmark. The second and equally important metric is scalability.

Cognitive-X:

The mission at Cognitive-X is to deliver custom cloud platforms, built specifically for applied analytics and data science.

Our customers are typically working with outdated platforms for their connected products and implementing data science or analytics to their products is virtually impossible due to legacy data storage and capture methodologies.

So Jim Heising, our technology visionary, architects super-light weight, server-less cloud platforms, designed to ingest and process huge amounts of data in an analytics-friendly environment.

The AWS based platforms are assembled using off-the-shelf AWS tools, and built specifically to enable applied analytics to the valuable data that is currently not being used effectively, nor is it generating any revenues.

The typical platform ends up costing our customers less than $300 per month in recurring AWS charges, despite handling well over 7bn data points flowing through it.

Building a Cognitive Layer:

Once the customer is setup with an analytics-friendly cloud platform, they can turn their attention to developing and implementing valuable analytics tools.

Deploying predictive analyzers and data signature pattern recognition is simplified and the innovation can continue unimpeded.

This is super exciting considering the rate at which devices and environments are getting “connected.”

And the price of sensors continues to plummet as the economic benefits of Moore’s law unfolds, and the viability of deploying a foundational sensory network inside of any given indoor environment is fast approaching.

Sensors are already deployed in IOT/connected devices, but they are in silo’s, specific to the devices/products and in most cases, they fail. Being in IOT hardware is a difficult sector these days, but that is a separate article.

With the right sensors, anything is possible.

Lessons from the Ecovent Founders:

The concept for enabling customers to be able to rapidly deploy analytics to their products came from my experience working with Ecovent technology.

The Ecovent tech is particularly interesting as they were the basis for the science behind the technology that I acquired while running ConnectM.

What made these brilliant founders special (Dip and Nick particularly) was that they based this system and its algorithms in applied science (physics and thermodynamics) of the ideal gas law; or PV=nRT.

Their technology was (and still is) ahead of the curve in terms of IOT and data science. In fact, their technology had a cognitive layer built into the product. By measuring pressure, temperature and humidity from 15–20+ locations throughout a building, the system could understand the thermal environment.

Understand Environment 1st, Then Command/Control:

Once the environment is understood by the system, the command/control function of the vents simply open and close relative to the setpoints for each room, providing precise control of the climate for each room.

Although this technology is still super interesting, it was the cognitive layer and the application of science to sensor readings/data that was most appealing from a business case and scale perspective for us (Jim, Conor & I) as the founders of Cognitive-X.

Affordable Sensory Network:

In fact, if you leave the vents out of the system, with about 15 sensors (each with pressure, temperature and humidity), and a hub/gateway (smart things, or similar), you can likely outfit your home or the floor of any building for ~$150.00, and likely a cloud subscription of ~$4.99/month. This is affordable to most, and has a basis for a business model in building out a cognitive layer if there are measurable returns…

Once the sensory network is established, the Cognitive layer can be applied, and this is where the data becomes important. The sensors provide the data, but unless there is interpretation of the data, it is meaningless.

Side Note: Check out Synthetic Sensors:

*This gets even more exciting when you consider the “Synthetic Sensors” that are coming to market, lead by Gierad Laput (@gierad) Yang Zhang and Chris Harrison and team.

With a cognitive layer of data interpretation, the environment becomes visible and understood.

With visibility, there is awareness of real-time conditions, and with awareness of real-time conditions, actions can be taken in a way that adds value and saves money. Similar to the controllable vents, but the applications are essentially limitless.

There are countless applications here, but our focus is on what presents a business case and is scalable.

Order of Magnitude Business Case:

Even better is a business case that provides an order of magnitude impact to value and savings, as this is where breakthroughs occur (Reference to Peter Theil’s “Zero to One”).

Occupancy sensing for lighting saves electricity and has a business case if the cost is offset by sufficient and measurable savings to justify the investment…but is it at an order of magnitude? I don’t think so.

Organic Growth Condition Monitoring (Mold Condition Monitoring) however does appear to offer an order of magnitude business case in terms of value…well, particularly in areas where mold growth is a known problem.

The comp or benchmark today is ~$300-$500 for a one-time mold assessment of your home. Then there is the removal cost which can get into the thousands.

But mold and other organic growth, only grows if the conditions are right. The conditions are directly related to indoor relative humidity and temperature. So, if you can monitor the conditions and receive alerts when the threshold is exceeded for a specific amount of time, you can take corrective action…i.e. turn on the air conditioner, or de-humidifier.

Back to the business case — Assuming you have deployed temperature and humidity sensors (estimated cost ~$100), the data analyzer we developed can provide monitoring and visibility to these conditions for less than $0.10 per analysis to the end user.

If you scheduled an analysis every other day (183x per year), this would cost the end customer $18 per year…or $1.60 per month. Compare that to one-time assessments of $300-$500 and you have an order of magnitude business case, providing more (183 assessments vs. 1 assessment per year), for less ($18 per year vs. $300-$500 per year).

The Cognitive Layer Potential:

This is where order of magnitude ROI becomes possible by leveraging sensor data and applying the cognitive layer. The cognitive layer alone provides the business case, and once in place, command/control functions become simple to put into place. I.e. Use an IF THIS, THEN THAT function — IF Mold Analyzer alerts, THEN turn on AC.

This is the starting point for Cognitive-X — starting with order of magnitude ROI business cases with our customers and work our way back to develop the platform and tech to make this happen at scale.

While convenience IOT products like Google Home and ECHO are certainly valuable, we will leave those to Amazon and Google…we will be focusing our global team on solving the stuff with an undeniable ROI and value of a new order.