According to a PwC analysis, global GDP will be up to 14% higher in 2030 as a result of the accelerating development and take-up of AI – the equivalent of an additional $15.7 trillion. The economic impact of AI will be driven by productivity gains from businesses automating processes, productivity gains from businesses augmenting their existing labor force with AI technologies and increased consumer demand resulting from the availability of personalized and AI-enhanced products and services.
AI will transform almost all industries and business
Imagine you understood your customers so well that you could anticipate their needs and ship them goods even before they ordered them. Thanks to AI we are almost there. The possibilities are endless.
Let’s look at a few existing use cases of AI implementation.
You have had a minor car accident. You take a picture of your cracked windshield or bumper and send it to your insurance company. AI will analyze and classify claims from the submitted photo, responding to your clients rapidly. This can be taken one step further; AI can be used to investigate potential fraudulent claims.
At the bank, AI can take client retention to a new level. Let AI analyze your vast client data and look for dynamic patterns of unhappy customers, so you can prevent them leaving. You can even let AI scan social networks to look for early signs of complaints and proactively mitigate their impact.
Your internal audit department often has too much data to be analyzed. AI can learn what the typical data and behavior of users or client is and give you early warning of potential frauds.
Use your AI to understand your clients and help them to achieve their goals. They only want your loan, because they want to buy a house and a loan is the only means to this goal.
Do you want to provide your retail customers or SMBs with personalized financial planning? Let AI look at their behavior, expenditures, income and help them to plan better, benchmark them with peers, and offer them help in the form of consumer credit or an overdraft.
Your call center is running overtime? Use AI to power your virtual assistant and answer questions and direct customers to the information they need.
Are you having nightmares with machinery failures and subsequent product quality degradation at your factory? Use AI to measure and analyze product specifications and machinery and you will be able to move from unplanned outages to planned maintenance.
There are many more already existing examples of AI use in real life. In retail, medicine, pharmaceutical research, cybersecurity, financial industry, shared services, automotive, in nearly every sector.
AI is not magic
But what is it? There are many definitions of AI, from those in popular sci-fi movies to scientific and IT definitions. One very pragmatic definition is: “AI is a prediction technology”. It does not give us much intelligence, as yet, but rather it helps us predict better. Predictions have been here for some time, so what has changed? As with electricity, light, cars, internet, and computers, there has been a fall in cost. Thanks to computer advancement, the adaption of prediction has been driven significantly higher beyond the areas of expected use, and costs have fallen.
Unlike ten years ago, we are not operating with a fixed set of parameters anymore. Now, thanks to advances in neural networks mathematics, AI can even take into account factors we do not know about and do not understand. This improves prediction accuracy, but also creates challenges regarding transparency and our ability to explain the result.
We use it every day but AI is not always understandable
You may not know it, but you are already using AI. It’s in your phone when it tells you how long it is going to take you to get home from work. It’s in your music application when you select your personalized playlist. It’s behind Siri, Alexa, Cortana and Google assistant. But again, there is not much intelligence here. The results of all these examples are predictions. Where you are most likely to travel after work. What you are most likely going to listen to. What answer to your question you are most likely expecting.
One of the challenges today is to understand how AI makes decisions. You want your AI to be always controllable, if not always understandable. Many AI algorithms are beyond human comprehension and we are not able to understand the reason for many AI decisions. This does not have to always be a problem. If an online shop suggests to a customer to buy a new T-shirt, the risk is low. But when a bank approves or declines a loan for a customer, you will want to understand how it came to that particular conclusion.
Where do I start?
You know AI is being widely adopted and you also want to leverage the benefits. Where do you start?
Firstly, look at your business from an AI perspective and identify where there is competitive pressure and operational problems. Then, prioritize areas of adoption and decide on your goals. Do you want to experiment with AI? Do you want to support your existing processes? Or do you want to transform your business? As you saw in the examples, AI can help you personalize your products and services, save time, enhance quality, predict supply and demand and much more.
And lastly, develop strategy and governance to drive the initiative and control AI.
Tomáš Martinák, Senior Manager, PwC