How to Assess AI Maturity
AI maturity is highly relevant to the successful development and execution of your AI vision and strategy. AI maturity relates to both data maturity (having advanced pipelines and infrastructure in place) and analytics maturity (having the ability to apply advanced analytics to both new and existing data). Having a grasp of your organization’s AI maturity enables you to make sound decisions with regard to risk and reward as they pertain to your AI effort.
A Technical Maturity Model
Before discussing AI maturity, let’s discuss technical maturity in general. I see technical maturity as a combination of three factors as they relate to a given technology or technical field at a given point of time (see Figure 1):
Experience: The collective amount of experience that a team has with a specific technical field or technology
Technical sophistication: A measure of a team’s ability to use advanced tools and techniques that relate to the technical field or technology
Technical competency: A measure of a team’s ability to successfully deliver on initiatives and projects that relate to the technical field or technology
Figure 1. Technical Maturity Model
As you might guess, higher levels of experience, technical sophistication, and/or technical competency increase technical maturity. Increased technical maturity boosts certainty and confidence. This, in turn, results in better outcomes and success.
An AI Maturity Model
I’ve developed an AI maturity model that builds on the technical sophistication piece of the technical maturity model. This model is shown in Figure 2.
Figure 2. AI Maturity Model
As shown in Figure 2, the starting point of this AI maturity model involves building a data foundation—in other words, identifying, acquiring, and preparing the “right” data. From there, as indicated by the dark arrows, AI maturity reflects a progression in technical sophistication from traditional analytics and statistics, to more sophisticated modeling and experimentation, to advanced analytics techniques, including AI and machine learning. (Note that it’s not necessary to gain complete competency at every stage of maturity before moving to the next.) The red arrows indicate that the output at each stage of the progression can influence both the stages that come after and those that come before. For example, all three analytics-specific stages can output data that can be integrated into the data foundation. Similarly, outcomes resulting from AI or machine learning applications can (and should) be understood using traditional analytics and statistics.
Putting It All Together
Figure 3 shows the relationship between technical maturity (or more specifically, AI maturity) and risk and reward in any AI effort. Here, reward relates to differentiation and competitive advantage. As for risk, or uncertainty, it pertains to the following:
Time: How long the AI effort will take
Cost: How much the AI effort will cost
Performance: In this case, the accuracy of the predictive model under development
Requirements: The data, features, and techniques required to achieve target performance
Figure 3. Innovation Uncertainty Risk versus Reward Model - A risk versus reward model that accounts for technical (or more specifically AI) maturity.
This model accounts for three levels of AI maturity:
Been there, done that: This level of AI maturity results from having the requisite experience, sophistication, and competency to effectively erase any uncertainty and resulting risk. There’s a chance, though, that at this level of AI maturity, the techniques and technology used and resulting outcomes might be commoditized—meaning you might generate minimal rewards with respect to differentiation and competitive advantage.
Branching out, expanding horizons: Here, AI maturity is “in progress.” Not surprisingly, AI efforts for companies at this level of AI maturity usually involve greater risk because they require more exploration and experimentation, and, by extension, yield less predictable outcomes. But, they offer greater potential for differentiation and competitive advantage.
High risk, high reward: This reflects a low level of AI maturity. Companies at this level of AI maturity must gamble—venturing into the great unknown, pushing boundaries and pioneering the use of emerging and state-of-the-art technologies, but knowing they might generate massive rewards.
One caveat: This model suggests that a low-risk scenario always involves commoditized technologies and techniques and results in low reward (and vice versa). That’s not actually the case. Some tech giants, like Google and Amazon, are so far ahead of the general market in terms of AI maturity that even their “been there, done that” techniques and technologies are incredibly advanced—in other words, they’re nowhere near commoditized. Indeed, you might argue that these types of companies flip this model on its head, in that they reap high rewards with relatively low risk. All this is to say that this model is relative, and applies more to “regular” companies.
Conclusion
Understanding your company’s level of AI maturity—an amalgam of data maturity and analytics maturity—can help you make sound decisions with regard to risk and reward, and is critical to the successful development and execution of your AI vision and strategy. To learn more, check out my book, AI for People and Business.
