If you have worked around government, you’ve probably seen a lot of bad technology. This is criminal, because technology is supposed to make things better, faster, or cheaper. In government, technology sometimes seems to make a bad process worse even faster. To highlight this I want to juxtapose the government’s typical options in a technology deployment and a possible alternative that I highlight in the Case Study section of Microslices.
A recent client of my company was helping a defense organization that needed to fill a security gap. In the wake of large-scale data leaks and growing concerns about protecting sensitive data, this organization realized that some of its own lapses were a result of inadequate security manager training.
To remedy the problem, the organization needed to index and “rate” its security positions to assign a level of training for each of these individuals. Simply reading, analyzing, and interpreting thousands of security positions would take nearly 2,500 human labor hours. With 10,000 descriptions, this could require $1.6 million to complete the work just for phase one. A clear cut case for technology, right? Well, if you’re the government you often have three options for a problem like this:
- Reluctantly devote valuable time and resources to complete the task manually.
- Hire a consultant or contractor to find a new approach to accomplish the task.
- Engage in a years-long IT procurement process that, if you’re lucky, will get a solution in place in a year or two, but it will likely be obsolete as soon as it lands on your desk.
Together with Big Sky and a small analytics firm called Taste Analytics, the organization investigated a fourth option: The Microslices approach, which uses quick deployment to automate the work that would otherwise be done by consultants or contractors.
To solve the security position description problem, Taste Analytics’ engine ingests a set of keywords that represent security rating levels. The engine then trains itself to look for those keywords in a description and spits out appropriate rating levels along with a confidence level about the rating. As users provide feedback, AI kicks in and the tool teaches itself to do a better job. The more data the tool ingests, the more accurate it becomes. This approach addresses the security issue, but also frees up significant time and resources for more critical activities.
So how does this demonstrate Microslices? The partnership between Big Sky and Taste Analytics allowed the organization to use a highly specialized tool that was well-suited for the specific task at hand. Through automation, the tool was able to dramatically improve a function that, until recently, we believed only humans could perform. Together, these factors allowed the organization to achieve meaningful results on a significantly compressed timeline.