If you’ve ever watched Iron Man movies, you’ll notice Tony Stark’s different artificial intelligence (AI) partners. From helping him create a new element to helping him do chores and run tests, AI in the Marvel Universe is extremely advanced and helpful.
So how much of this is fiction and how much is possible? Let’s discuss AI within ERP software. While we may not be able to create the same machines as Tony Stark, we can use his fictional creations as a basis for some of the things we’d like to see in the future.
AI as part ERP software: What is it like?
AI in an ERP context functions as a basic version of that futuristic “sentient being.” Its practical applications include machine learning and natural language processing among other things. AI can recognize patterns and automate tasks in a blink of an eye.
It is being used to enhance the customer experience and to facilitate advanced inventory management. It is beginning to become part of cybersecurity. It is revolutionizing predictive analytics. A best friend to your marketing department, it can offer credible marketing insights.
As Malin Huffman, Senior Director of Product Management for Oracle – NetSuite put it, “By having a complete picture of our customers’ operational business data, we can provide better intelligence to unlock the hidden potential of your data and help you manage your business more effectively.” AI within ERP software is built to adapt to human input and to prioritize search results based on a user’s interaction history.
Much like how Iron Man can ask for the most ideal flight patterns for the day, AI within ERP software will allow you to ask for just about anything. As interesting as present-day AI is, what will the future hold, and what are the risks?
Is AI as great as I think it is?
Unfortunately, AI isn’t close to what we want it to be – even among the top ERP systems. If we could utilize AI the same way that fictional characters do, we’d be living in a very different world.
The introduction of AI has been a mishmash of positive and negative. AI machine learning capabilities can become less effective once a certain amount of data has been captured. Jobs have been lost with the implementation of AI, and as AI becomes more widespread, more jobs will be eliminated or redirected. It’s difficult to picture a friendly robot retraining employees to help them learn new tasks.
With the introduction of artificial intelligence also comes the question of ethics. As the development of self-driving vehicles unfolds, consider the pros and cons. The trucking industry is estimated at $700 billion dollars and supports over three million drivers, which is not enough to handle current demand.
Consider this scenario: A self-driving delivery truck is in route. A huge object falls off a truck in front of it. There is a motorcycle to the left and another truck to the right. Given that AI is programmed by humans, the most likely scenario might be for the truck to swerve left (hitting the motorcyclist) because it’s the smallest object. These types of scenarios highlight the ethical conundrum of AI, and the connection to human thinking. Algorithmic biases exist sometimes originating from human biases.
The power of AI presents businesses with the responsibility to consider the impacts their technology can have when things don’t go as planned.
Looking to the future of AI
While there are many things to enhance or figure out, the future is generally bright for AI. Even though AI hasn’t fully “arrived,” software developers continue to push the boundaries of what it can do to produce potential business benefits. AI continues to make strides into deep learning (typically an advanced form of machine learning capable of nonlinear reasoning).
ERP products such as S/4HANA and Oracle Cloud are beginning to introduce more rudimentary machine learning functions such as automatically paying A/P invoices that fall within certain thresholds and automating other period-end close tasks. However, employing staff with AI experience to operate and deploy these solutions is a different challenge altogether.
While we may not have a Jarvis at our disposal, we very well may have something close to it within the next decade.
