If your organization provides a product or service – which applies to just about any business on the planet – you, too, can benefit from Artificial Intelligence (AI). While implementing AI may sound daunting, it doesn’t have to be complex or expensive. This article covers the basics of AI and looks at some easy-to-explore use cases.
What Is Artificial Intelligence
Artificial Intelligence - or “AI” – refers to machines programmed to simulate human behavior in various forms. Fundamentally, AI covers technology that allows machines - i.e., computers and software – to “think.” Obviously, no machine at this point - perhaps ever – is able to undertake complex thinking akin to what the human brain is capable of. However, AI is certainly capable of handling or mimicking simpler human tasks and behavior. And in some ways, AI is already ahead of us. At its most advanced level today, AI-powered machines have beaten Chess champions and can solve complex problems through learning and predictive analytics. Nevertheless, today’s AI remains in a very early stage of what the technology may eventually be able to accomplish. It mainly carries out narrowly-defined tasks, including analyzing and predicting user behavior in various fields, including healthcare, banking, travel, retail, and more. AI is rapidly evolving – and no one really knows just how intelligent machines will eventually be. Apocalyptic scenarios like the Terminator movies where human hubris leads to machines becoming self-aware and ultimately attempting to wipe out their makers - will remain science fiction. (Let’s hope so.) But AI is evolving at a dizzying pace, and machines will only continue to get better at mimicking human behavior – and becoming smarter and smarter.
For the purposes of this article, we are going to focus on AI on a relatively basic scale. Let’s look at some examples of how to implement the technology into business workflows through some specific use cases.
AI Use Cases
Let’s start by looking at airlines. We’ve all seen flight attendants selling duty free items on board international flights. Most airlines do it, presumably because it’s a good additional source of income. Can AI help increase sales? Absolutely! Here’s an example:
Let’s consider the data points the airline already has access to. Every airline knows exactly who’s on their plane. Now, purchases are made via credit cards, so, for travelers who fly frequently and use this service, there is an entire purchasing history. The first step is matching the data. Who on this plane has purchased which items?
Next, uncover simple patterns. Let’s say John Smith often buys perfume on his return flights. Since the airline knows John is interested in perfume, it could send him its current perfume offers through an app while he’s waiting to board the aircraft (highly personalized and highly timely). Maybe they could even pair that offer with another item he purchases occasionally, let’s say a 15% discount for a box of Lindt chocolates. Understanding preferences can go a long way and makes upselling more effective.
In the next iteration, John may even be able to make the purchase within the app, which already has his credit card information, while waiting to board. This not only would be convenient for the passenger, but limiting or phasing out in-flight shopping by allowing the passengers to complete their purchases pre-boarding actually helps the airlines save money as well: The fully-stocked in-flight trolleys are heavy. Getting rid of them saves fuel.
Highly personalized and timely offers can help increase sales. Airlines today already have the information to make that happen, yet they aren’t connecting the dots. There is nothing sophisticated about this. A simple AI algorithm and a new “Offers” feature within the app will do the trick.
Now, let’s look at a different industry: banking.
Financial services have even more data on their customers. Why not help them be more fiscally responsible by applying AI to customer behavior and trends? AI could enable advanced customer-service functionality. For example, banks could use AI to analyze trends and alert customers that they are spending too much on restaurants considering the money coming into the account. Or that their electricity bill has been going up for the last three months and that they should check with their utility provider to see what’s going on. While not everyone would welcome such a service, a lot of people would. Just think of all the college students who are living alone for the first time and must learn to manage their money. Such alerts can help them become fiscally responsible and more valuable customers down the line.
Start with “easy” and slowly increase the complexity…when you are ready
The examples above show that AI doesn’t have to be complex or impossibly expensive. Of course, there are extremely sophisticated, complex, and expensive applications of AI. But that’s generally not how you start with AI. Most organizations that use advanced, complex AI today started small.
As exemplified above, you may already have the data needed to start implementing AI for your business. Each customer leaves a data trail, so all you need to do is follow the breadcrumbs. Of course, at this stage you’ll need a data scientist who understands computer science, machine learning, and deep learning to follow and make sense of those breadcrumbs—and spot opportunities. But you certainly don’t need an army of expensive experts. Not if you start small.
Personalized shopping list
Let’s look at another example of AI-implementation. This time, we’ll focus on grocery stores. Which kinds of data do they already have? Most supermarkets have membership cards and have thus already accumulated a vast amount of data on their users, including, notably, what each member tends to buy from the store. So how can AI help the store take advantage of this data?.
While mundane, managing a grocery shopping list can be a challenge. Lack of time, forgetfulness, distractions, etc. may cause customers to miss items they actually need. Why not help them by delivering a little extra AI-driven convenience?
In addition to membership cards, many grocery stores and chains have phone apps - and some of those apps include a shopping list feature. Other stores, of course, would have to start from scratch and actually build and launch a customer app that can run the shopping list feature. But what if we apply AI to the shopping list? Well, the AI algorithm would track and learn user behavior and preferences. So, if the user buys, say, two quarts of organic milk every Monday, it will automatically add two quarts of organic milk to the list on Monday. Now, the grocery store can also suggest alternatives, like a different brand of organic milk that may be on sale.
If something new or rarely purchased is on the list, let’s say steak, the app may suggest BBQ sauce or something else users with a similar profile put in their shopping cart when they buy steaks.
This is all relatively easy; no sophisticated algorithm is needed. But even such basic usage of AI, if implemented correctly, may help increase sales through personalized, relevant recommendations and offers.
Influencer marketing & personalized shopping list
Now, let’s look at how you can work with influencers and take the personalization of shopping lists a step further. Numerous influencers are on Instagram sharing their favorite dishes, which their followers eagerly reproduce at home. Let’s say you partner with these influencers and ask them to create a dish with certain products you seek to push out. Through a click, users can import the ingredients into their shopping list.
Let’s take it one step further: The app knows that some users are vegetarians, lactose intolerant, or have other dietary constraints or particular preferences. So it will automatically adapt the recipe with alternatives for their individual diet. There may also be an option for the cost-conscious shopper, one that adapts the recipe to have items on sale or favors less-expensive brands and products.
Through profile customization, shoppers could indicate observed holidays, dietary constraints, hobbies, professions, allowing the app to make more personalized suggestions. While some users may hesitate to provide too much information, there is a large user base who would welcome it.
This form of AI-implementation could work for many types of retailers. If done right, shopping apps can analyze data and recommend shopping list items with increasing accuracy based on user behavior and preferences.
Improved understanding of user preferences and behavior will also help stores predict what they need to order. With enough adoption and a system smart enough, stores will be able to order, well, smarter - and minimize shortages or overstocking.
In this article, we’ve introduced Artificial Intelligence and described some basic examples of how this type of technology can be woven into key business processes. In the next installment, we’ll delve deeper into AI-implementation, including an example of how EastBanc Technologies helped a client solve a particularly complex problem through AI and machine learning. Stay tuned.