It wasn’t all that long ago when Artificial Intelligence (AI) was considered futuretech. Today, AI and related technologies such as Machine Learning are used in countless applications. Still, we’re undoubtedly just beginning to discover what can be done.
But, what will it take to leverage the future of AI?
For answers, I turned to my good friend, Dr. Michael Wu. Dr. Wu has a Ph.D. from the University of California, Berkeley, in Biophysics and a triple B.Sc. degree in math, physics, and molecular & cell biology.
Dr. Wu is one of the world’s premier authorities on artificial intelligence (AI), machine learning (ML), data science, and behavioral economics. He’s currently the Chief AI Strategist at PROS, an AI-powered SaaS provider that helps companies monetize more efficiently in the digital economy. Prior to PROS (NYSE: PRO), Michael was the Chief Scientist at Lithium for a decade, where he focused on developing predictive and prescriptive algorithms to extract insights from social media big data. (See additional bio information below).
However, one of Dr. Wu’s greatest assets, in my opinion, is his ability to make the complex simple.
I hear many developers using the terms AI and Machine Learning to describe how their applications deliver value, but is this really AI we’re talking about?
Dr. Wu: I am hearing many people call technologies AI that aren’t really AI. Because it’s such a hot topic, everybody wants to be part of it. In some cases, the AI moniker helps companies recruit talent or attract funding. However, that doesn’t mean that all of the applications out there are inaccurately classified as AI.
One of the simplest examples of AI in use today is “Personalization AI.” Netflix, Amazon, and other online shopping sites use this branch of AI to recommend products based on how you’ve interacted with them in the past. Did you buy a product they recommended? Did you actually use it. For example, did you watch the movie on Netflix? What other products have you searched for? And what products have consumers like you searched for and bought? Using this data, they adjust the model to refine their recommendations for you.
Other types of AI are in use today as well. Cognitive AI, sometimes called Perceptual AI, includes chat bots and digital assistants that learn to use human language and “see” the world the way we do, so they can better fulfill our requests. Familiar examples of these include Siri, Alexa, Cortana, and Google Assistant.
Then there’s Autonomous AI, an AI class that drives autonomous machines. Self-driving cars are a consumer-oriented example, but there are many industrial ones, too. Amazon’s warehouses use this type of AI to run robotic carts that shuttle products around their fulfillment centers. They have yet to replicate the human dexterity of warehouse workers in fulfilling orders, but they are advancing rapidly by learning from their human colleagues.
For the insurance carrier, Decision AI is probably the most relevant category of AI right now. Its primary application is to help people make better decisions or to automate the decision-making process. Much of what we do at PROS is Decision AI. Setting prices involves intensive data analysis and numerous decisions. By using Machine Learning and AI, we help our customers define and refine their pricing models to set more optimal price points.
In an insurance scenario, you might have an agent on the phone with a potential buyer. You want to recommend the policy that’s right for the customer, but you also want to recommend one that the customer has a high likelihood of selecting in conjunction with many other criteria. With data gathered from that customer as well as from other customers, including data on how insurance products are used, not just purchased, AI could recommend a policy that will simultaneously optimize all the objectives in a consistent, data-driven fashion. And AI can analyze all the data that goes into this decision far faster than any human agent can.
What are the challenges of implementing AI like that?
Dr. Wu: One of the main challenges is the availability and accessibility to the data required to train the AI. Today, sales interactions are between people —the customer on one end of the line and the agent on the other—are not digitized. Even the best agent can only gather so much data while interacting with a customer, let alone remembering all the details. Carriers will need to look for ways to make the digitization of sales interactions more seamless, right down to what’s said, how it’s said, and how the customer responds.
Certain tools can do some of this now, but they’re just scratching the surface of what can be done. As consumers grow more accustomed to interacting with digital agents though, digitization should become easier
What types of talent should businesses be looking to hire if they want to implement AI?
Dr. Wu: We’re at an interesting transition right now. Businesses still have to hire AI professionals who can build or create these AI applications. They may even require data scientists to create the initial models, software engineers to develop the program stacks, and so forth. They have to respond to a kind of chicken-or-the-egg problem. In other words, the initial model has to be good enough that people will want to use it to allow the AI system to collect more user feedback and learn from that data. If people don’t use it, the AI application won’t have any feedback to learn from, and the model won’t improve.
In the future, implementing AI will be easier. Compare it to the computer industry. Fifty years ago, the only people who used a computer were computer engineers who wrote the code to perform the jobs the system was designed for. Fast forward to today, and computers are entirely user-friendly. Of course, you don’t need to be super technical to use one even though the typical smartphone today has as much computing power as the super computer of fifty years ago. The same will be true in the future for AI-based technologies.
What would you say to the business leader who, if not skeptical of AI, is at least in a “wait and see” mode?
Dr. Wu: I would tell them they’re risking their business. AI is already driving a whole new level of efficiency that any brute force effort of just throwing more people at the problem will never match. Once this technology becomes mainstream, it’s going to be difficult for an organization to play catch up since the technology will continuously evolve and build on itself.
*His research spans many areas, including customer experience, CRM, online influence, gamification, digital transformation, and AI. His R&D won him the recognition of Mark Zuckerberg, Marc Benioff, and other industry giants. Moreover, CRM Magazine acknowledged him as an Influential Leader. Michael believes in knowledge dissemination and speaks internationally at universities, conferences, and enterprises. His insights have inspired many global enterprises and are made accessible through two e-books aimed at the layperson: The Science of Social and The Science of Social 2.