AI use cases in operations

Depending on who is saying the word operations, it can have significantly different meanings. From a management perspective, it includes controlling Quality Assurance (QA) programs, hiring & training, and process monitoring & maintenance, all while creating strategies to enhance efficiency and production.
“Operations” is the fulcrum upon which productivity and efficiency are balanced. A large part of that process is obtaining the required raw materials, turning it into goods, and then getting it out the door, on the way to the consumer.
If you think that is all there is to it, however, you’re forgetting about having people that can complete each stage of the task; about obtaining feedback so you know the process is working; and, most importantly, coordinating the SRM, PRM, and CRM (Supplier Relationship Management, Production Relationship Management, and the Customer Relationship Management, respectively) factors.

The “other” sort of Operations

When some people say “operations,” they are very specifically referring to IT Operations. This is an incredibly complex area of business, too. Troubleshooting an IT problem at 4 AM from your bed with a laptop (because your boss called to say that “all credit card authorizations are failing”) is a nightmare.
Bosses traditionally think you “find the problem” and “fix it.” It’s a little more complicated when you’re looking at thousands of records that have been generated in that period, and then pawing through thousands of lines of code like a bear in a garbage dump, trying to find an obscure error.

AI to the Rescue!

Fortunately, both types of operations are amenable to being repaired with the same cure: Artificial Intelligence. Remember, AI’s chief feature is the ability to handle much more data than a human can in the same period of time. Examining thousands of records, looking for patterns, is ideal work for an AI.
It may be more straightforward, analyzing every logical step, but it is so fast that millions of decisions can be interpreted in a way that appears to emulate the human “intuitive leap.” By quickly eliminating all the nonsensical answers with its incredible speed, it finds solutions in seconds or minutes that could take a human hours or days to resolve.

More Than Raw Power

It also encompasses Machine Learning (ML) so that each favorable, neutral, or negative decision (and its result) is stored as an “experience.” All of these experiences are incorporated into a Neural Network (NN) which is like nothing so much as a relational database where similarities are connected.
Such an NN attaining the interconnectivity of a human brain is still a goal for the far future, well out of reach of the current technology. Our modern-day NNs have cross-associational paths which do allow them to make obscure connections, primitively, but similar to humans.
Typically (for example), a system designed for classifying job applicants can identify that a Geophysical Engineer and a “Prospector” have many overlapping talents. It will “know” that one shouldn’t necessarily be eliminated from a job search because the words chosen don’t match our expectations.

WATSON

IBM’s WATSON AI was brilliant at one thing when it was built: answering Jeopardy questions. It took thousands and thousands of programming hours and millions and millions of dollars to make it work. It was a true idiot savant, capable of doing one task perfectly, and very little else.
WATSON has evolved since then (at the behest of its creators—not by itself!) and is now more accurately described as a suite of programs with fairly wide-ranging analytical capabilities. You can use a free version on the internet for testing or can hire a version of your own (referred to as AIaaS, or Artificial Intelligence as a Service), with particular modules, to analyze your Big Data.

DeepMind

Even more sophisticated than WATSON, Google’s DeepMind uses reinforcement learning just like a human brain. WATSON was preprogrammed with data to be able to complete its function. DeepMind learns from its environment, experiences, and failures.
It requires no pre-programming. You turn it loose on a problem (such as a video game), and it bumbles around losing until it gains an insight with a “desired” result (say, an increase in score). Leave it overnight playing the game thousands or millions of times, and by morning it has become better than a human player.
This ability will allow it to analyze medical data to help doctors make better treatment decisions. It will help chefs build better recipes. It will reliably solve engineering problems, find new molecular combinations to create new drugs, or calculate how to deflect an asteroid on a collision course with Earth.

Planning, Organizing, Coordinating, Analyzing

Irrespective of which branch of operations you are in there is always scut work. We hand it off to people who are not in a position to complain that it is menial, trivial, or unrewarding. Up until now, the reality has been that we don’t have much choice anyway. The work must be done so that the whole process can move forward.
AIs, on the other hand, don’t care if work is seemingly trivial or menial. You don’t have to flatter a computer or pay it more, to get it to slog through the mire of line items, event logs, or daily reports. Now all you have to do is train your people in how to pose the right question to elicit the answer that you need.
You’ll still need the same people to understand which questions need to be asked; to analyze the answers to see if they make sense in context; to interrogate the AI and make sure that its reasoning was sound; and finally, to forward the results of the analysis to the original questioner.
Not only can you assure your people that the AI is not going to take their jobs, but they are going to need to expand their capabilities with additional training (at company expense). Your people already understand your systems, policies, corporate culture, and have established relationships with others in the company. Hiring new people that already have the skills means hiring neophytes and it will take months for them to get up to speed.

How can AI help me?

AIs will become an essential means to economic growth in the next decade. If you’re late to the table, expect scraps… AI will deliver better customer service, better control of MRP and ERP, fewer errors in every aspect of your operations, more efficient production, and greater profitability.
It will give you a level of Agility that has never been available to you before. This is due to the increased ability to analyze Big Data, predict trends, outthink your competition, and maximize your Business Intelligence (BI).
Equally important, all of the members of your team will be happier and more content with their jobs because the drudge work has been eliminated. Not only will any additional training they receive increase their loyalty towards the company, but now their time can be spent creatively applying their skills to find new and better ways to accomplish tasks. Your 20 or 30% churn rate might be reduced to single digits!

Catch Up Time

We have had a shortage of Data Scientists since the AI evolution began—there were just not enough to go around—attributable primarily to educational institutions not providing courses and C-suite executives reluctant to believe that the tech could advance so quickly.
Consequently, students were not willing to take that path with no guaranteed jobs. That is the C-suite’s job after all—to be conservative and make sure they don’t take a financial juggernaut down the wrong path.
Now that AI is so clearly defined as where we’re headed; students are studying Data Science in increasing numbers. They’ll be ready in the next three or four years, with larger and more abundant crops ready to be harvested in subsequent years.

Get your AI system in place

You may elect to hire your fleet of data scientists (at an average annual salary of 100,000 U.S. dollars each) or to have it offsite as AIaaS (making it cost-accessible even to small companies). Whichever you choose, it needs time to be integrated with your existing software and hardware stack, as well as all of your processes.
Faced with a lack of available labor as the unemployment rate continues to fall, AI offers a solution in automating our manufacturing. You’ll retain your current workforce to manage the AI manufacturing, but the lack of additional workers won’t cripple you.
It is a much better alternative than uprooting your manufacturing facility and placing it in a new location (or even a different country), with lower wages and a larger workforce. Operations, after all, is all about improving efficiency and productivity, and not about adding new capital costs that could persist for decades.

The Takeaway

Knowing whom to hire when you receive 500 résumés for each opening can be overwhelming. Creating new business by offering a one-off (unique) product designed precisely to specification could be an expensive mistake if all the subtleties are not understood. Dealing with thousands of questions per month from non-paying customers investigating a Free Trial offer can be even worse. Knowing what products are going to be popular this season, and which raw materials you will need order is often a shot-in-the-dark. Will political tensions, weather, or insurrection interfere with your supply lines?
All of these challenges and hundreds more can be solved with a properly deployed and trained AI. You now need AI to stay competitive, let alone to excel.
Please feel free to start a conversation with us here. Our plug and play platform is the ideal tool for end-to-end implementation of AI in your business.
Don’t worry because we will be there for your entire journey, not just a quick sale! We look forward to hearing from you and helping you get your business safely into the future, free of both Skynet and HAL9000!