Why Doesn’t The Chart Go Beyond Today?

One day AI systems will predict what is currently impossible, to allow us to overcome the inefficiencies in our lives and in our factories.

In a previous blog, I discussed the reasons why we humans do not trust artificial intelligence. In this blog, I would like to look at the potential of AI and what, without the human mistrust barrier, the world might look like for a manufacturer.

The Chart on the Wall

I was recently walking around a factory and stopped to look at all the graphs posted on the wall, productivity by line, number of accidents to date, yield by shift, cases per hour, overall equipment effectiveness (OEE). The wall was awash with color and every conceivable chart type.

Looking at the images got me thinking, “why don’t any of the charts go beyond today?” Knowing what is going to happen in the future not based on the past, but based on having more information and using AI, would be a game-changer!

What if I knew that five people will be sick on Monday, my supplier will be late on Thursday because of planned roadworks and my largest customer will cancel his order as he has a cash crisis?

If I could predict the number of accidents based on history, shouldn’t my staff be told that one of them will be hurt this week? I know it’s Monday and production is always five percent down on Mondays as the plant needs to stabilize after the weekend. I know five percent of employees will not perform as well because the home team is playing away on a Sunday. I know all of these things, but I don’t do anything about them because they may be wrong. There is still an element of risk to my calculations since they are based on a limited set of data and my ability to assimilate it.

What if I had more data than I can compute in my brain? What if I could overlay every dimension of the factory and external influences and compute the whole as a single system? If that were so I could predict the future.

Predictive Analytics

Predictive analytics is the process of using data mining, statistics and modeling to make predictions about future outcomes. In other words, historical data defines a set of parameters, which computers can then use to determine what user behavior/responses might be in the future, add to this artificial intelligence, data from external sources and bingo you could predict tomorrow!

What would that mean for the future? Well, it would mean that we will all have to think and manage our lives very differently.

Take a simple example. An operations manager looks at what has happened last week to be able to create his charts, he has the data, the yield, labor sickness rates, downtime, number of late supplies, it’s all there and easy to assimilate and make sense of why something in the past happened. He probably knew the outcome before he started and was looking for the reasons.

What would happen if he could also look at what is going to happen tomorrow, now the operations manager is presented with what is going to happen next week, it’s not a guess anymore. It’s not just statistics based on history, it’s predictive information based on an artificial intelligence model that has learned to be right.

He now knows that on Wednesday line five will break down, five percent of the staff will not turn in on Tuesday, Supplier A will get stuck in traffic making deliveries late, and one person between Tuesday and Friday will have an accident, but it won’t be serious.

Given this knowledge, the operations manager would have the ability to mitigate against everything that is going to happen and plan for it.

Science Fiction?

One may say that this is pure science fiction, but it is not, given enough disparate pieces of information we can all predict the future in our daily lives.

When attending a meeting, we consider every piece of information we have on waking up, showering, eating breakfast, petrol in the tank, traffic jams, parking availability, the time of day and the effect of the weather.

We use our experience for some of these like how long it takes to prepare and eat breakfast, how long it takes to park but for others where we have no experience we rely on data. Data from various sources, Google travel time, busy periods in garages, weather forecast, traffic predictions on roadworks and closures.

If we gave an artificial intelligence model for all of that information, it would out-think us and give us an exact leaving time, it will be interesting to see how we deal with it. We will probably still add on an hour, just in case!

To close a quote from a forward thinker, Steve Jobs:

“You can’t connect the dots looking forward; you can only connect them looking backwards. So you have to trust that the dots will somehow connect in your future. You have to trust in something – your gut, destiny, life, karma, whatever.” 

Perhaps Steve Jobs should have replaced “whatever” with “artificial intelligence” and then we may be able to make the chart go beyond today.

If you enjoyed this blog please look out for others on IFSBlogs.

I welcome comments on this or any other topic concerning process manufacturing.

Connect, discuss, and explore using any of the following means:

• Twitter: @ElkinsColin
• Email: elkins@ifsworld.com
• Blog: http://blog.ifs.com/author/colin-elkins
• LinkedIn: https://www.linkedin.com/in/colinelkins

Follow us on social media for the latest blog posts, industry and IFS news!

LinkedIn | Twitter | Facebook

Photographer: Kristopher Roller