The advent of the computer age has transformed manufacturing and business productivity has soared.
Evolution of network availability and efficiency have enabled the creation of new applications in all walks of life and are now being leveraged to explore new avenues to increase productivity.
One of these avenues is the Internet of Things (IoT).
The ability to capture data via connections to the Internet has created Artificial Intelligence (AI) through Machine Learning (ML). Machine learning enables “recognition” of patterns in data, as a machine “learns” how to make adjustments to an industrial process and increased production without the input of labor.
Production facilities benefit from operating completely with autonomous machines, and beyond the cost of real estate and environmental regulations, there is little cost difference between producing in the US versus offshore.
Machine Learning: the Differentiator
A more significant step forward for the economy is how ML enables smaller businesses with specialized labor resources to scale far beyond their traditional capacity to scale a new or proprietary process many times over without any additional human physical involvement. Ultimately, the limiting factors become the boundaries of technical ideas and innovation.
At this early stage of ML, the critical issue for companies is how to recognize what data they need to capture that will allow them to differentiate themselves over their competitors and thereby scale their business. The closer the decision-maker to the application data, the better he will recognize the extractable value of that data, and understand how it will drive the business forward.
The next major challenge is then how to gather that data efficiently and economically, compile it into a manageable format and then enable self-service analytics to extract usable conclusions from the available data.
The Importance of Architecture
Selection of the initial appropriate architecture is hugely important. If the application design requires streaming technology, the sheer quantities of data will be enormous, and the continuity of that data is critical. While it is perfect for many applications, the public cloud is not suitable for these types of application. Its cost model will become prohibitive, as storage demands increase, and network latency caused by poor proximity to the data source cannot be overcome. Equally important, is selection of a software platform appropriate to the application rather than adapting an application to software available from the public cloud vendor. Changing the application architecture down the road will be hugely expensive and hugely disruptive. Once ML is in operation there is no turning back.
Advances in technology and available tools have opened up opportunities to capture available data and analyze it to gather insights into efficiencies and behaviors, create new products and services through those insights, and thereby redefine the economy. Machines will expand the reach of human ingenuity through artificial intelligence and enable smaller, more nimble organizations to effectively compete with large, more ponderous ones. The danger lies in management not recognizing how AI will impact the business, and how to avoid making poor architectural decisions early in the process.