The use of Artificial Intelligence in supply chain data analysis
With the latest update of Microsoft 365, we have all been given the functionality of A.I. in the form of their latest version of system assistant, Copilot.
Arguably, it is simply a reflection of how A.I. has become increasingly dominant in our daily lives, but some see it as the start of a dystopian future which is so often imagined in the many Sci-fi films using the trope “human vs machine”. Will robotic A.I. take our jobs? Replace the Soldiers, Firefighters and Police officers? Replace medical staff and Teachers?
Whilst A.I. replacing certain careers requires an unholy amount of suspension of disbelief, one area where it is feeling far more plausible is in the Business Information sector. With packages like Power BI, A.I. is already making some creative suggestions, and so it would not seem such a stretch to imagine all data analysts being replaced by computers which are driven by A.I. in the fullness of time.
To a certain extent, this could even be beneficial in some facets of the job: certainly in the handling of large swathes of, say, printed matter such as delivery notes or invoices, the idea that after (or during) the laborious scanning process, the A.I. could actually look at those scans, and extract key information and deposit it into a spreadsheet whilst you get yourself a coffee, seems appealing. And whilst pretty much all companies manage at least some of their business transactions on computers nowadays so those days of tedious manual scanning or data entry are far less common, it is concerning how many of those companies do still manually enter their data with no control over the field data entered.
Recently a number of client data records had to be “cleaned” because they had irregularities of column headings and information fields were not precisely captured, and this was fairly tedious work. However, from a desk, there is no better way to understand a business than in the forced study of its details because a steady perusal of data and subsequent careful recreation or tidy up of the information can be enlightening. It can show patterns of misunderstanding within the client’s employees, or highlight anomalies, or show errors unknowingly carried over for months or years in some cases, and one gains an insightful understanding of the business as a result of the work.
With understanding and insight, coupled with experience and imagination, improvements in many different aspects of the business can come to light. Could A.I. do some of this cleaning? No doubt. But could A.I. draw inferences, and conclusions, and suggest changes based on the findings? No. Perhaps A.I. could do the data entry more accurately in less time than a tired human who is struggling with their concentration span on any given day, but the data would be cold, hard facts. To interpret them and offer fresh insight or innovation, would need a human.
Computers are, at their core, binary thinkers. Yes/no; black/white; there is no nuance. Yes, they are programmed using higher level computer languages but ultimately every command ends up in binary and it is this which is executed. There is no emotion; no perception, no creative thinking; no intuition. Merely a logical rendering of everything in front of it, according to the precise set of rules, parameters and criteria laid out in the operating system or program commands.
It must be acknowledged that a computer is far faster and accurate than most human brains at executing these logical, programmed tasks – look how quickly pocket calculators (and subsequently, calculators on phones) replaced mental arithmetic and/or scribbling on notepads, so these sorts of functions can, and are, given over to the computer and rightly so. This capability, whether harnessed on desktops, watches or phones and tablets, is absolutely here to stay and is enthusiastically embraced by most, and very few [if any] people, feel threatened by it. But the other stuff? That requires a certain type of thinker – one with creative and emotional intelligence, and the capacity to infer, deduce and innovate. Artificial Intelligence is the branch of computer programming and software development where the programmers are trying to instil these very characteristics into the hardware we use daily and it is getting better all the time, but it has a way to go yet.
Should we be threatened by A.I.?
With precise and specific instructions, verbal or written, a computer can be commanded to carry out logical and discrete actions and it will do so, as fast as its multi-core processors and clock speed are able – and this is getting faster all the time, meaning that our technology paradoxically now ages faster than ever as it gets outstripped so quickly. IF a computer is capable of “going off code” or, taking previous answers and applying new, different ways to process or interpret those answers, we move towards true A.I. but, and it is a huge “but”, surely this would mean that the computer would have to be able to essentially code itself with new command lines without the need for a computer programmer? The main difference between a human brain, and an incredibly fast computer – aside from coffee – is this very thing, the capacity for Neuroplasticity, neurogenesis and synaptic reorganisation. The brain can “rewire itself” by new experiences and new habits for example, whilst a computer cannot. It can only be reprogrammed: self-programming? Not so much. Perhaps by looking for blocks of code that it can discern for itself would do the job, by running the code in a “sand pit” application before taking the code into its own core processes, a computer could mimic neuroplasticity, but again, as always, this new code would have to already exist somewhere.
Where will A.I. Add the most value?
A.I. generated work is as a rule, plagiarised – by definition, it can only return results and use information that it already knows about and therefore, which already exists: someone else has already thought of it. In our field, if we start relying on A.I. to redesign our warehouses or supply chains, or control our inventory, it would have to apply the methodology, parameters, and criteria of the closest match of a business in the public domain that resembles yours but by this very definition, the solution will never be bespoke; will never fit you quite as well as a perfectly tailored suit would do. Their stock, their customers, their local road network, their values, their aims, their budget… it will never be exactly the same as yours and therefore an A.I. solution will never match up to one created by humans who are really good at what they do.
Undeniably, A.I. could probably beat someone who is perhaps inexperienced at what they do, so perhaps those people might have cause for alarm, but on the whole, specialists are unlikely to be replaced by it, only supported by it. Back in the field of data analysis, yes; some mathematical or statistical models are used frequently and repetitively, and these would of course be perfect for A.I. generated output, but the interpretation of that output is where the skill truly lies and this is of course where the experience, and knowledge of an industry expert comes in.
Whilst many routine aspects of jobs – even including some types of surgery/medical treatment – are able to be programmed to strict criteria, and therefore carried out safely by computer driven robots which follow the rules and stay within certain parameters, there will always be a need for the oversight of a free thinking, intelligent, learning, and experienced human in case the unexpected happens.
At ASCALi, we do not believe that any of us will or should ever be completely replaced by A.I. – but we must learn to work alongside it. AI will prove as useful a business tool in our supply chain design, implementation and optimisation projects as other technological advancements over the last few decades such as CAD, Heuristic Modelling and Scheduling tools have been. The difference is, that it will become far more integrated to everything we do, and we are looking forward to embracing that change.
