What is AI, and why does it matter *now?
At its most expansive, AI is the capability of a machine to imitate intelligent human behavior. But what we think of as ‘artificial intelligence’ tends to change over time — once the capability becomes familiar and reliable, we consider it just plain old software.
In 1958, the idea of computers playing chess was AI; in 1998, recognizing a face in a photo was AI; today, they are standard software elements.
‘Practical’ AI
From a practical business perspective, the most valuable ‘intelligent human behavior’ that AI can replicate is the capacity to recognize patterns and create knowledge from data and then use this understanding to make profitable decisions.
Importantly, the digital capabilities required for machines to make such decisions are growing at breathtaking speed. Everybody knows Moore’s Law — crudely stated, that information processing productivity doubles every 18 months. The less well-known Kryder’s Law and Nielsen’s Law observe that digital data storage and transmission productivity also double every 12–24 months.
These extraordinary growth rates have been compounding for decades, driving innovations from the origins of the Internet to the explosion in Big Data processing. As a result, each year the feasibility frontier for digital applications advances and software engineers race to build tools that can exploit this burgeoning capacity.
Beyond just these technological advances, the nature and potential sources of data are also changing rapidly. Companies have always relied on data to make decisions, but the advent of systemic data availability, especially external sources of data not typically stored ‘in-house’ at a company, are starting to significantly enhance decision making potential. Weather forecasts, competitive marketing calendars, consumer preferences and local events are all examples of available external data sources that can be brought to bear on decisions by using AI tools to process them within the required time frames.
These collective advances have now reached a point where technology development is crossing the threshold of human cognition in specific areas. In plain English, many useful business decisions can now be made by software — better, faster and more efficiently than by us humans. At the edge of the frontier, AI is creating immense business value by automating and optimizing business decisions that cannot be made reliably and cost-efficiently by people.
Typically, these decisions have 3 characteristics:
With the advancement of the feasibility frontier, examples of AI that can support these types of decisions are suddenly everywhere:
In each case, the pattern is the same. A company uses costly personnel for high- dollar decisions in a business process (e.g., major procurement events, ‘hero’ product images, major medical claims) but cannot cost-effectively deploy people of this quality for the much longer list of smaller decisions of the same type. Often the collective profit impact of these smaller decisions can be enormous — in some cases bigger in aggregate than the total of the large decisions.
By deploying AI systems, the corporation gets several benefits:
- Adding the incremental profit stream of human-like decision making for this larger set of decisions.
- Freeing up the expensive people working on the ‘edge cases’ to focus on the smaller number of highest-impact and most complex decisions.
- Providing additional low-cost support to the people making the highest impact decisions — i.e., ‘helping the B players perform more like A players’.
The team at Foundry has consistently observed that AI initiatives that make money for large corporations are much more prosaic than robots playing Jeopardy. They represent the application of data plus math to create tangible and measurable improvement in a repeated business decision process. And importantly, successful AI programs tend not to create these benefits through a single integrated transformation project, but rather through discrete initiatives focused on a series of individual processes. In essence, you build a ‘mountain of pebbles,’ and in a Global 2000 organization, each ‘pebble’ can be worth millions of dollars per year.
~ The Foundry team