This article originally appeared on Babson's Thought & Action publication.
Much has changed for artificial intelligence (AI) since 2011, when an unusual episode of “Jeopardy!” brought together two of the game show’s all-time winners to take on Watson, an IBM super computer.
Since that milestone moment, which saw Watson emerge victorious, AI has grown only more and more common, says G. Shankar, a Babson College professor of information technology management. Popping up in a host of disparate industries, AI has been used, among other things, for cooking, oncology treatment, inbound marketing, human resources, and global supply chain management.
Those hoping to implement AI and take advantage of those opportunities, however, may feel overwhelmed at first. “After all, implementing such a system is not inexpensive, and each system has a number of moving parts,” Shankar says.
For business leaders and entrepreneurs curious about AI, here are some tips for getting started with this ever-progressing technology.
Learn the Basics of AI
If you don’t know much about the basics of AI, the first thing to do is to educate yourself. So says Rich Palmer MBA’16, co-founder and chief technology officer at Gravyty, a Newton, Massachusetts, provider of AI-enabled fundraising software.
“There are so many free resources out there,” says Palmer. “They can help you speak the language of AI.” For instance, actor Robert Downey Jr. hosts a YouTube series called “The Age of AI,” and computer scientist and educator Andrew Ng provides a number of lectures on the topic.
As for Babson, it offers an online edX course, “AI for Leaders.” That course is taught, in part, by Tom Davenport, an AI expert and the President’s Distinguished Professor of Information Technology and Management.
Do You Even Need It?
AI may be a powerful tool, but Palmer says you shouldn’t assume that your business needs to invest in the technology, at least not yet.
He advises that you simply observe operations at first. “Start backward from observing people and their problems. AI should not be the default solution in your mind,” he says. “Starting with the solution before you fully know what the problem is can be a dangerous thing to do in the AI world.”
Talk to others in your industry, Palmer says. They are probably faced with similar issues. Did they invest in AI, or did they try other solutions?
No Moon Shots
When businesses decide to give AI a try, Davenport tells them not to become too ambitious. “Think big, start small,” he says. “Avoid moon shots. Go for low-hanging fruit.”
If you’re having trouble handling customer support inquiries, for instance, you may consider examining chatbots, a sort of virtual assistance, he says.
Palmer says there are now a growing number of companies, like his own Gravyty, which offer AI solutions that are tailored to specific industries. Whatever issue a business may decide to address with AI, it may not need to build that solution on its own. A tech venture may have already sprung up to do just that.
Just because you’ve invested in AI doesn’t mean your work is done. The hard part is data, says Palmer. Data that feeds into an AI system needs to be accurate, unbiased, and in the proper format.
Getting the data right, and fine-tuning the model associated with that data, is about 90 percent of the work with AI, estimates Palmer. “The sad truth with these AI algorithms is that the data has to be close to pristine,” he says.
Finally, remember that AI is progressing at a fast clip. “AI is a rapidly changing field,” says Palmer. “What you know today will be different tomorrow.”