If you’ve used the internet today, chances are you’ve recently interacted with several A.I. systems without even realizing it. Any time you browse an Amazon page, scroll through your Twitter feed, or run a Google search, you’re interacting with systems that rank content based on what these systems predict you’ll do next.
A large number of internet companies rely heavily on machine learning and A.I. to run their core operations. Large tech companies are some of the earliest and most prolific adopters of A.I. today. But as time passes, an increasing number of businesses will begin to incorporate A.I. into their processes, resulting in new efficiencies and optimizations in nearly every industry.
While it’s easy to claim that the use of A.I. will lead to major changes, it can be hard to pinpoint exactly what those changes might look like. By examining exactly what A.I. is and where it has already made an impact, we can begin to infer where it could go next.
Machine learning (ML) has already played a major role in the creation of new products and businesses. It now affects how consumers purchase products online, how companies manufacture goods, and even how we relax and enjoy our downtime.
Traditionally, machine learning has been exclusively used to identify patterns in data that can be used to make simple decisions or single predictions. For example, credit card companies have long used ML to identify fraudulent transactions, and insurance companies frequently use machine learning to make pricing decisions. Machine learning also determines which ads you see, what shows you’re recommended, and what products you buy.
But machine learning is also used heavily in non-consumer situations as well. In industrial environments, machine learning can help to pre-emptively fix components that have a high likelihood of failure, and today machine learning is used to optimize the operations of data centers. By examining how machine learning is used today, we can start to figure out how it will change business in the future.
Virtually every business manager has heard of A.I., but many of them are still deciding how to best apply it to their organizations. A large portion of them believe that they can use it to improve existing products and processes. In a 2018 survey conducted by Deloitte, the top two benefits of A.I. cited by business managers were enhancing existing products and improving internal operations.
AI’s leading benefits are enhanced products and processes— and better decisions
While we’ll undoubtedly see the use of A.I. result in additional process automation and improvements to existing products, we’ll also begin to see new markets and new products emerge from the most innovative companies. In addition, we’ll start to see companies that currently rely on human R&D and expertise begin to leverage A.I. for new product development and discovery. In the next few sections, we take a look at a variety of different sectors of the economy and how we expect to see A.I. applied to businesses in those industries.
The basic materials and industrials sectors aren’t areas of the economy that most people think about too often, but companies in these sectors form the foundation of many of the physical products that we buy and foods that we consume. While it’s easy to think of these as old and traditional businesses, companies in these industries are already making a push to incorporate A.I. techniques.
One of the most obvious ways that A.I. will make a difference in these industries lies in the area of preventative maintenance. Materials and industrial companies have thousands of production plants, components, and machinery, and can use decades of historical data to predict when this equipment is likely to fail. By proactively fixing these machines before they break, companies can avoid costly downtime and optimize how they spend money to purchase new hardware. Manufacturing companies will also use A.I. to help prevent repetitive motion or stress injuries in factory workers. Companies today are already beginning to experiment with smart wearables, and can use A.I. to analyze all of the data collected from the sensors within these wearables to identify workers that are at risk for injury.
The agriculture industry will also benefit from the use of A.I. Using satellite imagery and data collected from ground sensors, companies will begin to predict and optimize crop yields. They’ll also be able to use computer vision to identify individual plants, spraying herbicide only on weeds, rather than on an entire field.
Miners will begin to use A.I. to predict the best locations for new mines by using historical drilling and geological data. They’ll also be able to optimize how they collect and process ore samples by predicting which samples contain high concentrations of precious metals or valuable minerals.
Most of the companies that provide goods for everyday consumer use will begin to use A.I. to optimize sales by either creating highly targeted products or offerings, or to predict and identify trends before they become big.
The fashion and retail industries are beginning to incorporate A.I. into their sales, marketing, and operations efforts. Increasingly, stores like Amazon Go, where there are few or no employees, will begin to pop up in major metropolitan areas. Retailers will begin to increase the use of automated checkout and point-of-sale technologies. They’ll also use A.I. to identify the best location for new stores using their historical retail and geographical data.
Fashion retailers with an online presence will begin to identify and curate personalized outfits and clothing items based on who visits their sites, and clothing-in-a-box brands will begin to use A.I. to identify items that a customer is highly likely to purchase, supplementing their stylists’ recommendations.
While retailers will focus on optimizing operations, media companies will begin to use new techniques to actually create and augment content. Speech processing techniques will allow companies to automatically create captions for videos and descriptions for images, and new breakthroughs in deep learning will allow machines to automatically generate audiovisual media content.
Gaming companies will utilize better engines to create more enriching experiences. They’ll begin to use techniques in reinforcement learning to improve a machine’s ability to react to a human player and will also use A.I. to actually generate gaming environments. Integrations with VR and AR technologies will allow gamers to experience extremely immersive, constantly changing and adapting environments.
Most consumers will experience A.I. in transportation when new modes of travel become popular. Large automotive and technology companies are already working on self-driving cars, which utilize A.I. Looking a bit further out, we’ll see A.I.-controlled passenger-carrying drones and helicopters that allow urban and suburban residents to travel quickly through the air.
Oil and gas companies, which have been slower to adopt A.I. technologies today, will increase the use of A.I. to detect and deliver their product. Drillers will begin to use A.I. to detect underwater oil seeps and to conduct subsurface well data analysis, enabling them to find new sources of oil that would have otherwise gone undetected. And on the consumer side, computer vision will be used to identify when someone lights a cigarette too close to a gas pump, enabling operators to shut down pumps when there’s a safety hazard.
Healthcare, which comprises 18% of US GDP, and has a very large number of sub-fields, represents one of the most exciting areas for A.I. Nearly every area of healthcare can benefit from the predictive capabilities of A.I. Below are some of the salient use cases, though it’s extremely likely that as adoption of A.I. in healthcare increases, many new and interesting use cases will come to light.
In the field of radiology, experts examine medical images of patients to determine the cause of an ailment. Computer vision techniques are now being used to achieve expert-level results in radiology, and as new advancements in A.I. improve the ability for machines to generalize, it’s highly likely that A.I. will be used increasingly to supplement an expert’s diagnosis of a medical image.
Additionally, new non-invasive diagnostic procedures may be created which preclude the need to use blood draws or exploratory surgeries. For those ailments that present in external symptoms, computer vision and sensor analysis can be used to identify the root cause of a problem.
Techniques in A.I. and simulation can lead to the discovery of new molecules, which can act as the foundation of a new drug. By allowing A.I. systems to learn how existing chemicals and molecules affect the body, A.I. systems can begin to explore how new combinations of chemicals or new molecules can lead to new ways of impacting our health.
Additionally, A.I. techniques can be used to proactively identify adverse drug reactions on an individual basis. Today, the FDA, drug manufacturers, and other organizations collect vast amounts of data on unwanted or unexpected side effects from both individual and combined drug applications. By learning from this data, A.I. systems can identify which combinations of drugs may result in unwanted side effects in individuals. In a similar vein, these systems can examine historical data to understand where an existing drug might be useful for a new therapeutic application.
There have been several large breakthroughs in genetics in the past few decades. Geneticists are now able to identify and capture vast amounts of information from individuals. A.I. systems can learn how specific genes result in specific outcomes for different patients, which could then allow physicians to create highly customized treatment plans based on a patient’s individualized scenario.
A.I. can also assist with an area of research that attempts to predict how proteins will fold based on their chemical makeup. The shape of proteins, which are 3D structures of amino acid chains, can significantly influence many different areas of our health. Unfortunately, today, predicting how a protein might fold based on its components is very hard to do. By providing A.I. systems with enough data on known protein shapes and components, we can begin to understand new protein structures in the future, which could lead to in-depth preventative treatments and cures.
Finally, finance, an area of the economy which is accustomed to the adoption of new technologies, can benefit from the use of A.I. in both consumer and business applications across the back, middle, and front offices. Much like in the area of healthcare, applications of A.I. in finance are too broad to enumerate. However, we’ll outline some of the most common and immediate applications of A.I.
Capital markets are some of the most technology-forward in the world. High frequency trading, predictive analytics, and high volume transactions are key examples of how the world of investing relies on software and data management capabilities. As the use of A.I. becomes increasingly ubiquitous, buy-side and sell-side professionals will be among the first to use new A.I. techniques to optimize their work processes.
A.I. will be increasingly utilized in portfolio management, construction, and optimization. Companies will start to take into account an individual’s needs and provide highly targeted and personalized investment recommendations at scale.
Investment management firms will begin to use A.I. to identify and predict asset prices based on hundreds or thousands of variables. Portfolio managers will begin to automatically identify leading indicators and collections of datasets that are predictive of future price movements, and will be able to combine these predictions with simulations and statistical analysis to create profitable trading strategies. They’ll also be able to optimize how they route and execute orders by using new techniques in reinforcement, learning to find the lowest cost channels to purchase new assets.
Finally, outside of capital markets, private equity and VC investors will begin to leverage existing data to identify high potential deals earlier and with more frequency.
In the world of retail and consumer banking, individuals and banks will begin to see A.I. play more of a role in lending. New forms of credit scoring will emerge which take into account a user’s online and behavioral profile, allowing banks to more deeply understand consumers, and potentially enabling them to lend to a larger group of people without sacrificing risk. In fact, according to a survey by Baker & McKenzie, half of all respondents believe risk assessment will be the area of commercial banking that benefits the most from A.I. Finally, advancements in natural language processing will allow banks to more efficiently process the vast amounts of legal contracts that they generate every year, significantly reducing the number of loan service errors and hours required to service a loan.
While A.I. will lead to a future full of new possibilities, most businesses are still working to determine how to best use techniques in machine learning and predictive analytics in their organizations. Some of the most innovative organizations will move quickly to not only enhance their existing offerings, but to open up new markets and provide category-changing products. However, most existing businesses will incorporate A.I. in phases, applying it in targeted ways that will result in operational efficiencies.
Firms that move quickly will reap the benefits of providing market-leading products and interfaces, however they will also need to solve the biggest challenges in their industries. Yet even as the most nimble and forward-thinking businesses incorporate existing A.I. techniques into their operations, researchers in academia and industry will continue to improve and create new breakthroughs in A.I.
Ultimately, the future is certainly bright for businesses that look to use A.I.
Shanif Dhanani is the co-founder & CEO of Apteo. Prior to Apteo, Shanif was a data scientist and software engineer at Twitter, and prior to that he was the lead engineer and head of analytics at TapCommerce, a NYC-based ad tech startup acquired by Twitter. He has a passion for all things data and analytics, loves adventure traveling, and generally loves living in New York City.