By Sarah Nicastro, Creator, Future of Field Service
Perusing tech news headlines or looking at conference agendas, you see AI everywhere. While you innately know better, with the messages being delivered, it can be easy to be persuaded into thinking, “I bet AI would solve all of my problems.” We had Seth Earley, author of The Artificial Intelligence Powered Enterprise, on the podcast not long ago to discuss why he feels AI has failed to deliver on its promise to businesses everywhere and his advice for how derive value from the technology.
The AI Conundrum
Seth points to a few reasons he feels AI is failing to deliver on its promise. The first is what I eluded to in the intro – the hype surrounding the technology. “One of the biggest challenges we have with AI and machine learning is the tremendous amount of hype in the marketplace. AI is an umbrella term and the technologies we're seeing today have a history of components and underlying algorithms that have really been around for decades,” he says. “We see anything with an algorithm being called AI and what that does is it creates the wrong expectations.”
According to Seth, this hype leads organizations to look at AI as something that is brand new which feeds unrealistic expectations. “I think the big problem is that organizations are looking at AI as something that's brand new, and that it is going to solve problems that they haven't been able to solve,” he says. “In some cases that may be true, but it also creates unrealistic expectations. That's partly because whenever there's a big shift in technology it creates uncertainty about what this means to the business. This leads people to start making investments without really understanding the capabilities of the technology, or what processes they needed to really address.”
Seth isn’t saying AI isn’t valuable, but rather that companies need to look beyond the hype into the realities of not only the capabilities of the technology but, more importantly, what the business case is for their company for AI. “We see a lot of money being wasted on AI projects that are trying to boil the ocean or solve very intractable problems,” he says. “I think the biggest challenge is that many organizations don't have the basic foundational elements in place. Foundational processes, or the quality data that they need, and they're not necessarily understanding the nature of the problem they're trying to solve before going down this path.”
Advice for AI Success
To derive value from AI, Seth offers some advice to help organizations avoid the hype and unrealistic expectations and look at the technology through the lens of how it can solve business pain points. The first step is to know what those paint points are – to know what problem you are trying to solve. “The foundational piece is understanding what problem you're working on solving. And then, looking very carefully at the technology and saying, "What can this technology do, realistically?" he says. “There's a lot of successes out there, but there's a lot of failures. The reason for those failures really has to do with a lack of understanding of the true capabilities of the tools, and the processes that people are trying to enhance, and the business outcome that they're looking for. And, of course, not having the data.”
With the hype surrounding it, AI can be portrayed as a superpower-like technology – but the reality is, it isn’t magic. Like all technologies, for it to work for your business requires a strong foundation. “You can’t automate something you don't understand, and you can't automate a mess,” says Seth. “While AI is powerful, you still have to teach the technology about your business – your product, your services, your solutions, your customers. You have to give it the terminology that you're using, and the concepts that are important to the business. An ontology is basically a framework for that. I think the key piece for leaders to understand is that this is not magic, and there's a lot of foundational work that needs to be in place to make AI work. It's not sexy, it's really the basic blocking and tackling. You still have hard work to do; governance is important, and metrics.”
Part of this important foundational work is knowing the needs of your customers. “A lot of organizations fall down by not necessarily understanding the needs of their customers,” says Seth. For instance, we worked with a company that wanted to do personalization for their customers. So, we built the architecture, had some algorithms. And then, at the end of the day, they couldn't define what the personalized content should be. They couldn't say, well, how is this audience different than this audience? What do they need? They didn't know. They didn't know enough about the customer in order to use the technology to personalize that experience.”
Seth suggests starting small and building upon success. “Starting off with big, ambitious goals that stretch the organization and stretch the technology is inherently risky,” Seth explains. “That doesn't mean you can't have a big vision of what AI can do for the organization but going through the process of planning and doing small experiments will yield a lot of value. These experiments deliver learning and maturity that needs to be built up in order to be successful.”