September 25, 2017

Enterprise IoT: AI, Analytics, and Aviation

In part 1, we reflected on a discussion about IoT security in relation to the enterprise. The next part of our conversation focused on the changes in data collection and analytics, as well as the impact on the aviation industry. Our four guests span a diverse range of industry experts:

“I think one of the most exciting parts of AI and ML is when you can start to use it at the edge to make less explicit decisions about a thing or occurrence happening.”

-Allison Clift-Jennings, Filament

Some  of the most intriguing developments in the IoT are being done in the area of data collection and analytics. There's so much data that's being collected, much of which is simply generating a baseline understanding. When anomalies occur however - that is when this data starts to become actionable. Allison Clift-Jennings, co-founder and CEO of Filament, believes AI and analytics will greatly impact how we interpret data from the IoT.

Most people associate AI and ML as working in the cloud and on very large data sets. But with edge computing, the actual device at the edge of the network contains enough computational and networking capabilities to make fairly informed decisions on its own. The cloud will still be incredibly important for larger insights, but what’s exciting about edge computing is that it’s starting to explore what it would look like to have machine learning dynamically create routes between devices and optimizing network capabilities.

“The [end user] needs contextualized, relevant information presented at the right time and in the right fashion and that requires machine learning, analytics, and big data platforms to move IoT forward. It's not optional. It's part of where IoT is going.”

-Thierry Sender, Verizon

Verizon’s Thierry Sender brought up the point that between enterprise, consumer, network, managed solutions and so forth, we are about to experience exponential growth in the number of intelligent devices that are sending data to networks. But in order for us to derive insights and learnings, all this data will need to be contextualized.

Thierry went on to discuss ROI and IoT, and how it’s not just about reporting back and receiving the information real-time, but rather how we use the real-time data for preventive measures. Including things from security attacks to a failure of a jet engine, applying real-time data learnings can address the need to proactively notify the right individual with the highest priority item at the right time and suppress the other notices. For Thierry, this is where the real value for IoT and machine learning come together.

“The exchange of data between all these different devices has to be mandated in some way, which requires not just context but also discovery...[In order] for IoT to be effective, it will have to go cross-enterprise, cross-industry - it won't just be information that's gathered.“

-Raj Singh, JetBlue Technology Ventures

Raj Singh, managing director at JetBlue Technology Ventures, uses the aviation industry as a great example of how ML in IoT can be helpful. Pilots are bombarded with a barrage of notifications while in the cockpit and are constantly having to work out which ones are important and which ones need to be suppressed.  It’s the same issue with the sheer volume of alerts, notifications, and information that the IoT generates. It's not sufficient to simply have a set of notifications. The critical function is in the sorting and filtering.

Due to regulations and certifications required in the aviation industry, the use of smart technology can be stifled. Everything moves a bit more slowly. For example, edge computing may certainly be valuable, but getting edge boxes certified on a plane takes two years because you have to be absolutely sure that it works properly. Though it’s a fairly different business model compared to more traditional IoT applications, the aviation industry is actively seeking out the technologies to start catching them up.

“At JetBlue we have 230 planes, 20,000 people, and millions of parts. Getting the right part with the right person with the right plane in the right place at the right time is a nontrivial problem. Technology can help us with that, especially technology related to enterprise IoT around asset tracking and understanding the health of those parts.”

-Raj Singh, JetBlue Technology Ventures

The aviation industry has woken up to the fact that they can use enterprise IoT to solve a number of their challenges.  They’re primarily focused on asset tracking (especially for luggage), predictive maintenance, regionalization, and other broader changes in transportation. JetBlue Technology Ventures keeps an eye on these changes, such as the rise of the autonomous car or Hyperloop or even the high-speed rail and the way it might impact the industry.

They are also keeping a keen eye on where the revenue is going. Aviation is a very low-margin business and the revenue for an air ticket is split among many different players. Airlines are looking to recapture some of that margin by going directly to their customers. But for somebody like JetBlue, a relatively small player with 5% of air business in the U.S., they have to think about how to convert people with their digital assets. Raj believes what will be crucial to their success going forward is establishing themselves as a place where people go for more than just aviation. They need to become a hub for all things travel.

Tags: Techstars , Filament , Verizon , JetBlue Technology Ventures , Enterprise IoT