'The robots are coming' - when will ERP AI become feasible?
While there is a considerable amount of market buzz relating to Artificial Intelligence (AI) these days, does the hyperbole amount to a real operational transformation, or does today’s noise just end up representing a ‘nothingburger’ when it comes to creating an ERP value?
To answer that question let’s first consider what AI actually is:
According to Merriam-Webster ‘Artificial Intelligence’ is defined as:
“A branch of computer science dealing with the simulation of intelligent (human) behavior in computers; or the capability of a machine to imitate intelligent human behavior.”
In this context, then, systems that tout ‘ERP AI’ may be just a couple of bridges too far when it comes to operational legitimacy, since in nearly all cases it appears that little or no simulated ‘human behavior’ is really involved.
Instead, today’s complex ERP systems leverage hosts of manually-developed scripts that, in turn, connect and interact with clusters of databases to identify, catalog, and index other data-sets, that become ‘information’ repositories for the user.
ERP scripting, not ‘ERP AI’
However, if an enterprise script developer makes an error when defining or identifying linear data requirements regarding a particular code effort, that failure can cascade through the system, leading to malformed results or worse. Consequently, while this activity ultimately may end up ‘suggesting intelligence’, or more rightly, a ‘simulation of a simulation of human intelligence based on human behavior’, it just ain’t ‘Artificial Intelligence’ at all.
This clear lack of understanding can cause all kinds of problems when it comes to dealing with enterprise expectation management, since until someone comes up with a 21st Century version of the HAL9000, ‘someone’ is going to be putting ‘someone’s’ hands on a keyboard ‘somewhere’. Once that happens; the creation of scripts; that interact with other scripts; that touch, and interact with other local and remote databases; will always be operationally limited by manual competencies that may, or may not, be legitimate.
Now for the good news
Now; all that said, there are a host of new capabilities regarding machine-to-machine information development, and the creation, and/or the execution of operational tasks.
For example, cloud ERP operations have fostered and encouraged:
- flexible innovations regarding automated scripts for business intelligence
- direct and indirect control systems operation, in the case of CnC machinery security monitoring mobile operations
- datastore management
- utility updating
- large-scale email processing
...just to name a few of today’s ‘automated’ values for ERP. However, in the main, these advantages are created by automated processes not ‘artificial intelligence’ in the truest sense.
This is not to say that at some point in the future, you won’t be able to say; ‘ERP System X, identify and catalog the number of widgets sold by Division Y for the last 20 quarters’; and the system responds with immediate paper read-out of the requested information. However, we’re still a bit down the road yet, and with miles of potholes to be traversed before we get there.
In the meantime, we’ll just have to continue to employ the old school way, meaning we’ll have to use our own brains to resolve our own problems, leveraging the systems we have at hand. After all, the point of computing in the first place is to support and increase efficiencies related to the human effort, not supplant it entirely.
ERP and cryptocurrencies - where do we stand?
How can we expect cryptocurrencies to impact ERP in the future?
ERP: what can we expect in 2018?
Why you should be keeping an eye on machine learning as one of this year’s most prominent technol...
Three of the most important ERP developments in 2017
2017's standout developments in the ERP world - any you want to add?