We need to be assured that our AI and analytic strategy should be as close to the business as we mature with various situations and assets
A battle between the Human Mind and AI-Analytics
Recently, an incident happened in the Australian Open Men’s Single Final during the Daniil Medvedev and Rafael Nadal match, which was circulated widely through WhatsApp. Medvedev, also known as the dark horse, was leading the first two sets and suddenly a message flashed on the live TV which was being viewed by millions of people across the world that Medvedev had 90 percent chances of winning the match and Nadal, only 10 percent. The message started rolling when Medvedev was already won the two sets and was leading the third set. But, in the end, Nadal won the match 2-6, 6-7, 6-4, 6-4, 7-5.
So that is what AI is.
AI is a technology that majorly depends on historical data and is only able to extrapolate within that boundary. It cannot go into uncharted territory. Until there is a well-structured data set, even an unstructured but readable one, AI and analytics will bring some projections & predictions, and some theoretical assumptions on what can happen under certain circumstances. However, if you ask the machine to provide you with the information on situations that never happened in the past, I think it is anyone’s guess what the answer would be!
Technologies like AI and analytics are widely dependent on the historical and incremental data set. As long as these models are fed with multiple variables and scenarios, they will become mature. But in any uncharted territory where a situation had not been experienced in the past, AI or analytics models will find it difficult to articulate or express the outcome. Here, we need to be assured that our AI and analytic strategy should be as close to the business as we mature with various situations and assets.
The conclusion to make here is that the human mind at any point can beat AI and analytics in any uncharted territory even if we have data and confirmed predictions. And this goes reverse as AI does not have any control over the human mind.
Manufacturing Industry and AI
Artificial Intelligence is most applied in manufacturing to improve overall equipment efficiency (OEE) and first-pass yield in production. Over time, manufacturers can use AI to increase uptime, improve quality and consistency, which allows for better forecasting.
The first requirement in AI or analytical models is how contextual data can be collated, gathered, and put under the respective models and how we can fine-tune that given the ground reality outputs or the results under various situations. So over time, we need to continue to capture the data, live & real. We also need to capture the resultant output under various conditions and feed it into the model to make it much more mature, realistic, and reliable.
The approach should be to automatically bind relevant models’ traceability data to the outbound machine or sensor data, in real time as it is read from the source, and store it fully cataloged as such in a scalable analytics database.
This is like - What if you had a system in place that automatically detected production issues in real-time before they happen?
The benefits would be predictive maintenance, inventory, and product outlier detection in an accessible and intuitive way, driving operational excellence to new levels.
Today’s Technology and Data collection
Technology has improved and matured a lot in the last three to five years. Today we are using technologies that can help us to gather data from machines that are fifteen-twenty years old.
With such advancement, we are now living in the pool of data, which in turn forces us to understand the data collection protocols, how to collect data safely and securely. As more and more machine data are getting integrated, the risk of losing data sets to external agents has also increased. Further, this also increases the risk of getting hacked and intruded, impacting the production.
The importance of data security and information security has also become a relevant and close-knit topic today. So, while the technology is maturing, we need to take control of where the data is collected from, where it is kept/stored- on-premise or cloud, and security measures while it’s in motion or at rest. With all this, the risk management from an information security perspective becomes very critical in today’s world.
Different organizations have their own look, feel, target, and benefits definitions. Some organizations will look at improved quality, customer satisfaction, and improved consistency as an output of their AI and analytics strategy that will ultimately improve their profitability and bottom line.
Then some want to monitor the effectiveness of the AI and Analytics strategy at each KPI level, for example - am I consistently improving the quality of my products? Many would want to leave it to their operation team or leadership and don’t mind improvisation of just one percent at the bottom level.
So, the business benefits must be decided and designed at the beginning so that you can measure and monitor the effectiveness of the strategy. In today’s world, it’s never been free money. So every single penny that you invest in any of these technologies needs to be assessed and measured for appropriate business benefits.
Additionally, when we talk about any technology deployment, it cannot go without a financial strategy to give the business all the benefits over a specific period of time. Finally, choosing the appropriate technologies relevant to the business or the industry is critical. So, in my assessment, the major factors that can make an AI and Analytics strategy successful in any enterprise are- deciding on the KPI’s output, pre-requisites, and the right technology.
At the end
In conclusion, all the above-mentioned factors are applicable to any business, industry, and specifically on the business intention of the leadership on how much they depend on the technology to push their business. Further, after deploying any technology it becomes critical to verify the outputs with the historical results and check the benefits on the ground before trusting the strategy by closing our eyes.
In general, Business is never static, it is a dynamic process. Around the globe, parameters of business- market dynamics, equations, and requirements keep on changing. One model that was perfect two years back may have failed miserably during the pandemic. But one thing which helped us today from the past and which may be helpful in the future (especially in scenarios like Covid), is the Data. Lastly, we can only predict or control what will happen tomorrow within known / past scenarios. That may be protected by AI or analytics tools as long as we are collecting the data, we are making the future-ready for similar unforeseen circumstances. Because -Yes. Data is the new bacon. AI/ML is taking it to a new height!
The author is Joint President &CIO at HINDALCO