How Can You Ensure Your AI Driven Project is Successful?

August 09, 2018
Roman Stanek's picture
Founder & CEO
Roman Stanek is a passionate entrepreneur and industry thought leader with over 20 years of high-tech experience. His latest venture, GoodData, was founded in 2007 with the mission to disrupt the business intelligence space and monetize big data. Prior to GoodData, Roman was Founder and CEO of NetBeans, the leading Java development environment (acquired by Sun Microsystems in 1999) and Systinet, a leading SOA governance platform (acquired by Mercury Interactive, later Hewlett Packard, in 2006).

There’s no doubt that AI is a hot topic right now. Siri, Alexa, and Facebook’s facial recognition technology have brought AI to the masses, and those technologies are just the tip of the iceberg—AI developments are in the news constantly. Given how pervasive it is, it’s therefore understandable that AI is also front-of-mind for many companies who are looking to jump on board and add AI to their data products. From increased revenue and better identification of new opportunities to reduced costs or risk of errors, there’s almost no limit to how companies can use AI today.

So Why do AI Projects Fail?

While the possibilities are numerous, there's still a huge disconnect between fantasy and the reality of a successful implementation. Companies set their sights on how exciting it would be to introduce AI in some way, without thinking through the business problem they are trying to solve. And when they start with the technology instead of the outcome, they often find that AI didn’t really pay off as they expected it to. In fact, Gartner estimates that 85% of big data projects fail—not a statistic that inspires confidence. So what steps can you take to make it more likely that your AI project will be successful and that you’ll begin reaping the rewards? Greg Satell at the Harvard Business Review reached out to me to discuss this very issue and the struggles associated with making an AI project successful.

With AI, Define your Business Outcome First

In my conversation with Greg, we spoke about the need to figure out what you want to achieve with AI as the first step. This may sound fairly straightforward, but unfortunately it’s not a top priority for most companies. They get caught up in the hype of what they’ve heard AI is capable of doing and has done for other companies, yet they fail to consider a realistic application for AI at their own company. That’s a big problem because—unsurprisingly—people won’t use something if it’s not an improvement over the way they currently do things.

AI Adoption is the Key to ROI

I find that AI projects often start with trying to implement a particular technology or approach. Perhaps an executive has heard about another company who has done something cool with AI, and they’d like to recreate that experience. Sounds great, but the problem comes when we get to implementation. Though the idea might have sounded promising within the confines of a board room, it’s usually quickly discovered that not bringing in front-line managers and employees—those who truly understand the business processes that you are trying to improve—makes adoption rates low or even nonexistent. They may not even trust what they are seeing because they weren’t involved in the definition process or training of the AI models. All the time, effort, and money that went into introducing AI is then wasted.  

Find Success with your AI Projects

Addressing this first step of defining your business outcome is crucial if you hope to find success with your AI projects. It may not be a particularly exciting first step, and the temptation is certainly there to just throw AI at your business processes and see what happens. However, you’re far more likely to be successful if you take the time to find a specific problem or process that AI can address.  Once this is done, you can focus on identifying which specific tasks are best suited for AI to handle and on transitioning your human employees to focus on higher-value or strategic work—and then move on to actual implementation. Do the hard work first to set yourself on a path to success.