AI is a broad field with both theoretical and practical applications for corporations. Thanks to COVID-19, new implementations of AI in the workplace are being explored. —
By Jill Motley
Artificial Intelligence (AI) is loosely defined as the endeavor to make computer systems perform human-like tasks.
Of course, the common understanding of what is “human-like” and what should be considered purely computer behavior an ever-moving goalpost. As a result of this, the term “AI” is commonly used to refer to the current “cutting edge” of human-like computer operation.
For example, a computer program that plays chess might be called “an AI” out of tradition, but actually thinking of that program as an example of true artificial intelligence feels almost quaint.
So, the question becomes, “What is the cutting edge right now?”
The global COVID-19 pandemic has rendered it impractical for humans to perform a wide range of tasks that they would otherwise have been doing. Everything from site inspections to customer service positions are presently under lockdown or other restrictions. To fill the gap, AI development and implementation has been accelerated, especially in corporate applications.
For a bit of technical terminology, the range of tasks that a certain AI algorithm is intended to perform is referred to as its “domain.” The domain of a self-driving car’s AI domain, for instance, is driving. AlphaGo’s domain is the game of Go.
There are three domains that have been particularly relevant to businesses during the COVID-19 pandemic. They are natural human-like conversation, repetitive tasks, and market analysis/prediction. The following sections will explore how these domains have been impacted by COVID-19 and how they are shifting the business landscape.
Talking to Robots: The Future of AI-Driven Customer Experience
Perhaps the most predictable leap forward in AI spurred on by COVID-19 is in the domain of conversation.
The necessity of social distancing has made it far more difficult for employees to communicate with customers. In-person conversations are less common, but many companies have reported overloads at their call centers. With widespread staffing shortages, filling the gap with human employees has not been possible.
Unfortunately, the average customer is more in need of information than ever, and their only means of getting it is over the phone or on the Internet. If they cannot connect, they will take their business elsewhere.
“Propensity models” are used to determine which consumers are most likely to purchase from a particular company, and which products they are most likely to buy.
This business imperative is why the chatbot and natural language processing industries have seen so much business lately. The day when an AI will be able to win Turing’s “imitation game” is still some time away, but both text- and voice-based conversational AIs are becoming more and more indistinguishable from human salespeople every day.
In fact, AI is not just used in the role of a boots-on-the-ground retail employee. It can also work directly with high-level marketing executives.
“Propensity models” are used to determine which consumers are most likely to purchase from a particular company, and which products they are most likely to buy. This data covers a wide range of demographics and can be time-consuming to parse, but if it is understood well, it can allow a sales team to focus their energy where it will produce the most return.
COVID-19 has radically shifted the purchasing habits of virtually everyone. The old, well-studied data no longer applies, and the new data has to be analyzed as quickly as possible in order to realize its greatest benefits.
Luckily, the development of a propensity model boils down to an exercise in spotting patterns within data. That is a task that AI researchers have been working on almost since the term “artificial intelligence” was coined. In this way, the benefits of AI data analysis are clear, and the only question is which companies will most effectively seize this opportunity.
Let the Bot Do the Menial Work
AI may soon be able to complete many tedious, unpopular jobs that currently can only be performed by humans.
Businesses have long relied on humans’ naturally keen ability to identify inconsistencies and details. For example, any institution that needs to check its client’s identification documents most likely uses humans for that task. Only humans have historically been able to identify the subtle marks of a fake ID.
Now, ID verification may be the crossroads of several AI subfields: facial recognition, optical character recognition, and other types of image analysis. AI may also be uniquely suited to completing other repetitive visual recognition tasks, freeing up the humans for more fulfilling, creative work.
Predicting the Future Using the Past
With the economic hardships caused by COVID-19, many manufacturers and retailers have found themselves with severely limited supplies. As shortages gain customer-facing visibility, panic and anxiety compete with loss of potential market share as front-running problems.
Considering these tough trade-off, those companies must put extra effort into shipping their materials and products where they are needed when they are needed. It is not enough to be “just in time” now, not with mail delays and shipping hiccups becoming normal. Rather, they need to get them there slightly before they are needed.
So All That They Need to Do is Predict The Future. Easy.
In all seriousness, AI is exceptionally good at this kind of pattern recognition and extrapolation. There have been bots trading on the stock market for years now, and the domain of predicting human economic behavior has already seen significant development.
Computer systems that use machine learning and AI to predict demand and extrapolate optimal resource allocation are called cognitive supply chains.
The difficulty with this AI domain, however, is that the global economy in its current form has never seen a crisis of this scale and nature before. There is no data that an AI system can use to reliably predict behaviors in these circumstances, so businesses are investing in systems with the capacity to learn on the go and process supply chain data with the efficiency of a high frequency trading bot.
Will it be enough? Time will tell. As chatbots, visual screening robots, and “cognitive supply chain” systems take hold, human productivity can be expanded. AI can help ease businesses through the current crisis, and firms will have to see – once the world reopens – which systems are truly worthy of being kept.