A.I. Island 3: A Case study in A.I. The Digital Transformation from a products and financial services firm to a digital industry.

    Published on

Ken Finnegan

Follow Following Unfollow Ken Finnegan

Sign in to follow this author

Chief Technologist – IDA Ireland Technology & TechIreland Advisor

Previously on A.I. Island I provided a simple introduction on how Artificial Intelligence will impact some jobs by augmenting the worker so they can become more effective at their job and an example of the type of jobs that are under threat e.g. highly repetitive jobs. ” A.I. Island 2: The impact of Artificial Intelligence on jobs. What does the future of employment look like?

According to IDC by 2019, 40 percent of all digital transformation initiatives will be supported by cognitive/AI capabilities. Many companies have begun the journey.

In this post I want to examine the digital transformation of a company, how they are impacted by this evolution in technology and the recognised benefits by undertaking the journey.

Engineering and Heavy Industry (AIoT) – A closer look at GE

Before I get into the case studies the diagram below is a simple example of the ‘why’ (benefits) and ‘what’ (motivations) companies expect from a digital transformation in current times. I will explore each of these topics in detail over the series of blogs but for now this is a simple articulation of the benefits.

When I was doing my MSc, GE was a company that was persistently used in multiple modules as case studies about management, leadership and innovation. I found an article during the summer that is a perfect example of a digital transformation journey and the innovative creative thinking the organisation undertook to become a ‘future ready’ company.

GE understood and recognised the value of data from an early stage. The company’s CDO, Bill Ruh, understood that being able to capture data from the organisations industrial devices, that he could get insights that would enable GE to capitalise on the benefits mentioned above. This is the real thinking about taking the first steps in becoming a data driven organisation, data is king.

By collecting real time information on mission critical systems they could create digital twins. A digital twin refers to a digital replica of physical assets, processes and systems that can be used for various purposes. The digital representation provides both the elements and the dynamics of how the device operates and lives throughout its life cycle. Jet engines, wind mills and gas turbines could be digitally represented and provide multiple layers of insight that were not previously available. Insights about the temperature, vibrations, noise and humidity could now be collected and stored in the cloud. This information created the twin with the physical asset and allowed for analysis to happen which enabled the creation of models that could identify potential issues with the assets before an event actually happened. This proved invaluable to GE, what they effectively created was a crystal ball in predicting the future health of their thousands upon thousands of assets in the field.

By harvesting massive amounts of data from their wind mills, they were able to identify the initial symptoms before failure that cause mills to fail and break e.g. a rise in temperature could represent friction from parts rubbing together, vibration data indicative of a loose part and even the data about sound of a mill could indicate blockage or resistance. Combine these pieces of data and compare them against mills that had previously failed and you know that you need to mobilise the maintenance team (or have the mill mobilise the team) – imagine this at scale; thousands of mills, tens of thousands of preventable downtime hours, better design ideas, a proactive maintenance team, enhanced reputation, new XaaS business models… the savings are phenomenal.

To do predictive maintenance, first we add sensors to the system that will monitor and collect data about its operations. Data for predictive maintenance is The former approach only provides a boolean answer, but can provide greater accuracy with less data. The latter needs more data although it provides more information about when the failure will happen. We will explore both of these approaches using the NASA engine failure dataset. time series data. Data includes a timestamp, a set of sensor readings collected at the same time as timestamps, and device identifiers. The goal of predictive maintenance is to predict at the time “t”, using the data up to that time, whether the equipment will fail in the near future. Predictive maintenance can be formulated in one of the two ways:

So right now you might be saying, this is just the IoT in action. And yes it is but include A.I. into the mix in the form of a machine learning approach. This make it possible to learn from new data and to modify predictive models over time. Ruh points out that machine learning makes it possible to identify anomalies, signatures and trends in machine performance and develop understanding of patterns of behavior. In addition, Machine Learning can be applied to help identify efficiencies within a machine and use this as a best practice for other machines, essentially benchmarking best performance. GE already has about 750,000 digital twins and is rapidly adding more.

I don’t want to get too technical in this blog post but here is a short example using Machine Learning Techniques for Predictive Maintenance for those interested.

There is an important message here. In order to effectively use A.I. or any component of it e.g. Machine Learning you need data, a lot of data. Data is the food, the very sustenance to keep A.I. alive or at least to get the best benefits. I recently saw the acronym AIoT. This really says it all in this example.

This simple model is representative of GE’s approach:

What I really enjoyed about this case study is the journey GE has taken in migrating from being an industrial and consumer products and financial services firm to a “digital industrial” company with a strong focus on the “Industrial Internet” and $7 billion in software sales in 2016.

Recently CIO Jim Fowler observed how GE has evolved into a “future ready” company, where “the technology is going to become the process” and where employees will work in “mission-based teams” that form to solve specific business problems and then disband to go and find and solve new business problems. DevOps in Business…

Chief Technologist – IDA Ireland Technology & TechIreland Advisor

Article by channel:

Read more articles tagged: Case Studies