AI needs to be a shared learning experience across the enterprise to succeed.
One of the lessons incessantly driven home for technology initiatives — particularly artificial intelligence — is the need to formulate a strategy that guides your technology vision and investments forward. However, to paraphrase management guru Peter Drucker’s famous phrase, when it comes to AI, lack of experience eats strategy for breakfast.
An AI strategy should only come into play as people in the enterprise understand its implications, urges Dr. Andrew Ng, globally recognized leader in AI. “I recommend a company start small, gain momentum,” Ng stated in a recent interview with CXOTalk’s Michael Krigsman. “Only after your company knows better what building AI feels like, then you’ll be a much more thoughtful place to try a thoughtful strategy.” Only after gaining knowledge and experience with AI can you gain “a more thoughtful vision for how AI will change your industry, how it changes where to play, where not to play, and what creates value.”
So, where to begin on an AI journey? What projects are suitable to get started? “When I work with a company, I usually recommend brainstorming at least half a dozen projects,” Ng relates. “Then for each of them, we spend a few weeks doing both technical diligence as well as business diligence, where we model out the bottom-line or top-line impacts. Only after we’ve convinced ourselves that this is a valuable project do we then commit resources to this.”
The challenge is identifying and moving forward use cases of value to businesses, Ng continues. “Over the last couple years, we’ve had a lot of companies run a lot of pilot projects that did not make into production,” he says. “It is difficult to set the right judgment about selecting the most valuable project to work on. I see a lot of startups running proofs-of-concept. We as an industry we still need to get better at both identifying the valuable proofs of concept, and then also making sure we take these things to production.”
Start with small, winnable projects to convince the rest of the enterprise of the value of pursuing AI, Ng continues. Even during his time at Google, Ng had to finesse the accuracy and business viability of an AI-driven speech recognition product before the larger organization began to embrace it, he reports. “My number-one advice to leaders is to start small, and then deliver a quick win maybe in six to 12 months. Then use that to teach the organization how AI works, what are available use cases, and do bigger and bigger projects over time.”