Back to top

Plan Ahead to Adapt Later

You can’t be surprised by change if you’re expecting it (and agilely ready for it)

One of my favorite stories of collaboration and tactical problem solving happened several years ago and about 250 miles above the Earth. The International Space Station had scheduled a space walk for two astronauts to repair a power system, only to discover a pair of problem bolts. In a follow up exercise to replace them, the astronauts found their solution by employing a modified toothbrush.

I love this story for a number of reasons. In one view, it highlights how a simple problem can become incredibly serious depending on the environment. If issues go unchecked or unexpected problems arise, you could find yourself in the middle of a crisis before you know it. Another point to note is all the tools the astronauts had access to aboard the International Space Station. Millions of dollars of hardware and advanced technology and what saved the day was a creative solution.

Despite the years of experience NASA has in space, the toothbrush project was not a pre-planned solution tucked neatly into a manual. What saved the day was collaboration across teams (literally off-world in this case), and the willingness to adapt to change in real-time. Of course, the stakes of solving a problem in real-time are not typically as critical as in this story, but the benefits can be just as significant.

In the case of artificial intelligence (AI) solutions, we always recommend that you start with small, clearly defined goals. There’s a few reasons to take this approach:

  • Your path to implementation (and results) is a matter of weeks instead of months or years
  • You’ll get really good at knowing what problems are best solved by an AI solution
  • You can easily iterate on your solution to grow out the responsibilities of your AI 

An AI can’t solve all of your problems overnight. Instead, plan to automate repetitive, high-volume tasks, get really good at those, and then build on your success.

The path to getting the most out of your AI solution is the same you’d take for building a company or training for a marathon. You start with small goals and work your way up to the bigger ones. In the technical space, this means you identify one problem, the ideal solution, and then build the process to solve it. A + B = C. If you want to get to ‘Z’ you have to go through the other twenty-five letters first.

If you try to solve for everything right out of the gate you’ll spend way too much time developing a solution, and most likely never get to implementation because it will be too complicated. Or you’ll take so much time and effort, that your solution will be outdated by the time you have a chance to use it. This is one reason why Shai was built with a flexible infrastructure, so she can operate in our customers environments, whether on premise, virtually, or in a secure cloud.

Let’s come back to another real-world example in the AI space. The woman who provided the speech for Siri first recorded her lines in 2005. Apple then debuted a beta Siri on an iPhone a full six years later. If you happened to interact with her, then you know there were plenty of bugs and she didn’t do a whole lot. Of course, the Siri of today is world’s better than that first version but that’s because she’s been updated countless times. In other words, Apple iterates on Siri to add more features and improve functionality over time.

By taking this approach (known as Agile in the IT world), you’ll be in a flexible position to adapt to changes, shift priorities, and solve problems at an incredible speed. There’s a reason kids learn to crawl and walk before they run—it makes for a better foundation to be successful.




Blog Author

By Brad Mascho
Chief Artificial Intelligence Officer, NCI