Five Questions for Getting Started with Machine Intelligence

"Getting Started," Chapter 3 of The Machine Intelligence Primer

Given all that you now know about MI, how can you get started on using MI to address problems you see at your organization, or even in your day-to-day life? First, it helps to consider what you’ve learned about the kinds of problems MI is and is not useful for today. 

As an emerging technology, the vendor, research, and expert ecosystem within which MI is developing becomes more complex by the day. It is difficult to discern which MI technologies and use cases will bring value to a problem space, and even more complicated to disentangle what MI implies for an organization’s future when it comes to its clients, teams, and long-term trajectory. As with any strategic initiative, the best place to begin is by developing a clear map of your goals, vision, risk tolerance, and readiness to deliver. 

Here's five key questions to ask that will help you get started using MI at your organization:

The first question to consider when evaluating MI investment is the value you hope to capture through the investment. With high expectations on the horizon, it’s important to clearly articulate the return on investment expected from MI in the near and long term. 

When considering investment in MI, it’s important to remember that commercial applications of MI are still nascent and no MI product or platform is truly “off-the-shelf” (yet). All but the most basic applications of MI come with a certain level of risk. It’s important that executives develop a clear articulation of how much risk the organization is willing to tolerate when it comes to MI so that they can set clear parameters with partners and avoid frustrating organizational leadership. 

MI experts are working hard to make machine learning less data intensive, but until then, MI requires a lot of labeled, well-organized data. This data intensiveness is a stumbling block for many organizations working to capture the value of MI. It is helpful to enter any MI engagement clear-eyed about the challenges to making use of data and consider strategies for addressing these challenges. 

MI talent is scarce, and the battle for the top experts is fierce. This can be quite disheartening for many organizations that can’t offer NFL salaries to talent who will only be with them for 12–24 months. But the scarcity of MI talent needn’t be a barrier to entry for MI investment. Through partnerships with academics and investment in existing talent, most organizations can build the bench needed to put MI to work. 

MI’s portrayal in the media can conjure up fears of wide-scale automation, job loss, and even robot violence. While many of these fears aren’t warranted (at least for now), MI does have the potential to significantly negatively impact our lives. It is therefore critical that executives consider how MI will impact their clients, shareholders, employees, and the public from an ethical, equity, and privacy standpoint when beginning any MI initiative. 

Download our complete guide to getting started with MI, which includes more detail and solution ideas to consider for each of the above questions, at the link below.