3963_InvisibleAI_Blog-1While companies are embracing AI, a recent study shows that their pursuits are driven more by fear of obsolescence than enthusiasm for change. Eighty percent of companies in a recent survey stated that they intend to expand their AI use over the next 12 months, but we also know that 77% of AI projects face challenges with business adoption, and less than 8% of companies have successfully implemented AI at the enterprise level.

Why so much failure? Poor adoption is a significant factor. In my experience, I see too much focus on the need for change and not enough on the people who will need to adopt it. On the other side of these changes are humans who need to buy into it, learn and then keep using these solutions successfully. These days, business transformation is happening at an uncomfortable pace and has left some feeling threatened by AI, so you need to make the adoption as easy as possible.

Adoption isn’t as difficult when AI is invisible. What do I mean by invisible? If you’re familiar with J.R.R. Tolkien’s fantasy novel The Hobbit, you might remember the main character’s sword, Sting, which glowed whenever goblins were near. While Bilbo’s sword was a necessity on his journey—enabling him to fight goblins—the sword’s magical element was not. It was an add-on that required no effort on Bilbo’s part and proved extremely handy. Invisible AI can work like Sting.

For example, one of my clients recently released an algorithm-driven suggestion engine to its sales force that recommends the next best customer for a rep to call on. Rather than create a new sales tool, they added the recommendations to a platform that the sales force was already using. This required no change management. It didn’t even require an internal communication. Sales reps noticed a recommendation in their CRM one day. This client could operate in stealth mode and simply watch reps adopt the advice. The more the reps used it, the better the advice became, and eventually adoption was very high. 

Invisible AI is about putting humans first and encouraging adoption as well as reducing fear and confusion. Here are three steps for pursuing invisible AI:

  1. Focus on your purpose. Going back to Middle Earth, Sting was created for a single, simple purpose: helping Bilbo, the user, avoid goblins. When you launch an AI project, start by determining what you’re looking to drive. Effectiveness? Efficiency? Productivity? Improved experience? Throughout your AI journey, keep your focus on these goals. Algorithms, technology and data sources can change, but focusing on your purpose must remain a constant.

  2. Shrink the change. As much as possible, try to minimize the change required to adopt your solution. Bring elements of AI to the tools that people are already using whenever possible to spare them a frustrating learning curve. With your focus on your purpose, look carefully at repetitive processes with well-defined, measurable outcomes and find the least obtrusive way to embed AI into those processes in a way that complements and aids human work rather than replaces it. For example, when one of my client’s products was approved for a new indication, they were able to dramatically increase intelligence around the product, market and competition by incorporating additional data sets and refreshing existing analytics models in their tool for tracking physician influence networks. This armed teams with the actionable insights they needed on the new indication during and after launch, without any learning curve or change management efforts. 

  3. Enhance the experience. Research on user experience and human interface design is relevant when embedding AI. We pursue AI for quantifiable benefits, such as ROI or measurable efficiency gains, but to achieve sustained adoption, it’s important to improve the user experience as well. Such improvements help reduce the skepticism and resistance that often accompany such a change. Bilbo’s sword glowed a subtle, pleasing blue that I’m sure he would have preferred over a flashing, garish yellow with accompanying noisy clangs. Be thoughtful in your design and, if possible, leave your users with a “wow” factor and a better user experience.

Bilbo undoubtedly recognized Sting’s usefulness at first sight. He didn’t need to learn any new skills (though swordsmanship would have come in handy), and he was understandably wowed by the appearance and utility of the blade. Similarly, when we roll out AI initiatives, we want our teams to instantly recognize its value, know exactly how to use it, and feel uplifted and supported by it. Pursue your AI transformation in this fashion and you’ll be much less likely to stumble during adoption.

Topics: user experience, artificial intelligence, data analytics, AI, analytics platforms, artificial intelligence & pharma, advanced analytics, UX