We haven’t come close to realizing the potential promised by Artificial Intelligence. According to Gartner’s Hype Cycle for Emerging Technologies 2018, which evaluates everything from Smart Fabrics to Blockchain to various AI tech including General and Edge AI and AI PaaS, only Deep Neural Nets (Deep Learning) has reached the “Peak of Inflated Expectations.” That is to say… all other AI tech in consideration is still undergoing significant innovation with years to go before a plateau.

What does this mean for you?

While Gartner isn’t the Alpha and Omega when it comes to tech, this is a pretty good bellwether for what innovative Chief Information Officers are doing. (Datanami unpacks it here). It’s as clear to analysts as it is to those of us entrenched in Machine Learning and AI that we have a long way to go. But opportunities for innovation have already arrived. Here’s how to benefit.

  1. Learn. Don’t feel like you have to be the “AI expert” on your team: the people driving this research literally have PhDs. Our competitive culture puts pressure on us to posture expertise, even if we’re swimming well out of our lane. To be clear, AI is not something you can self-teach in a month. But you can follow industry trends and learn where you might have opportunities to innovate your customer facing or internal products.

  2. Partner. Partnership gives you access to experts to guide you through the right decision-making and still set you up to be the AI expert in your organization. By virtue of working with experts in the field, you will intrinsically gain exposure to the type of information that you can’t glean from general articles on the subject. Your expert should be able to speak clearly and in an explainable way about how their models are going to work. While the algorithmic models are extremely technical, the problems they are trying to solve should be conceptually straightforward.

  3. Question. Be aware that not all machine learning models have full explainability yet. The “black box” problem – where we can’t fully explain why or how an AI system made the decision it made – is an industry-wide issue and is not yet solved. That’s why it’s a good idea to start small and test your models before fully integrating them into your systems. Starting small allows you to test hypotheses and implement proven methods.

Right now, for example, we have an engagement that has developed into a data exploration project. We’re helping an organization that couldn’t do deep analytics on their own, but we’re doing it in a very partnership-oriented approach where we’ve become a trusted advisor and are helping them learn how to eventually do it themselves.

Now is the optimal time to initiate innovation projects. “…by 2021, AI will support more than 80% of emerging technologies, and one year later will support more than 80% of enterprise IoT projects…”. If your company isn’t part of the surge, it will be left behind.

To read more about reducing the cost of (and risk to) implementing AI in last week’s blog, click here.