Processing: Enter AI & Data Science
Developing a spy mindset means understanding the five steps of the Intelligence Cycle.
1. Planning & Direction
2. Collection
3. Processing
4. Analysis & Production
5. Dissemination
Step three, Processing, includes sifting out relevant information from step two (Collection) and transforming it into meaningful reports for consumers to digest. Like movies on the silver screen, a lot of celluloid is often left on the editing room floor.
For imagery intelligence (IMINT), this means selecting the most recent photographs with the best views of the target – no cloud cover. For signals intelligence (SIGINT), this means selecting intercepts that address relevant issues – no idle chit chat. For human intelligence (HUMINT), this means selecting the source’s relevant observations – no detailed journal entries. Knowing what to ignore helps you collect what you need.
These raw intelligence reports, which are heavy on formatting but light on creativity – “Just the fact, ma’am” – usually contain text or images to convey the information, and are properly indexed to allow for electronic dissemination, storage, and retrieval.
The goal of Processing is to thread the needle – not reporting too little, not reporting too much – but the historical role of Processing was to create a body of raw intelligence reports that serve as metaphorical pieces of the puzzle for analysts to assemble into finished intelligence during step four of the Intelligence Cycle (Analysis & Production).
For this reason, Processing is often viewed as more meat grinder, less glamorous – more punctuation and elbow grease, less sex and danger.
However, things are changing. Processing has a bright future.
First, as the world becomes more digital, more information will be available in the public domain, some of which will be purchased for general use, not to answer specific questions. As such, one of the big challenges will be knowledge management: identifying, organizing, storing, and disseminating information within an organization.
However, terabytes of information are useful only if they can be transformed into products that convey meaning vis-à-vis your pending decisions. To succeed, the Processing step must grasp how the human mind processes information and have sufficient depth and complexity to convey meaningful information that reveals clear decision points.
Why rely only on human analysts to assemble pieces of the puzzle if artificial intelligence and data science can create puzzle frameworks for the raw intelligence to feed into directly and illuminate in real time?
In other words, the Processing and Analysis & Production steps may blend into one creative step, with the usual warnings of GIGO (garbage in, garbage out) and the fact that the puzzle frameworks are only as good as the people who design them.
For example, if sales are dropping, one way to assess the problem would be to overlay maps of where the sales teams are located with where most sales are being made, by your teams and by your competitors. One map could capture this problem intuitively, rather than flipping through pages of spreadsheets. Perhaps it’s time to update where the sales teams are physically located to increase time on target?
This is why the Planning & Direction step is so critical. Leadership at all levels should know in advance what their collection gaps are and what indicators they want to monitor daily, weekly, or monthly to ensure the ship is sailing in the right direction. There’s always space for bottom-up insights, but only to shape the Planning & Direction step.
Returning to the sales example, once the key variables or indicators for tracking sales are understood, the intelligence team can construct a dashboard that will integrate these steams of data into intuitive displays that remove human error. The boiling frog metaphor often causes people to modify their standards over time, such as saying you will sell a stock if it dips to $20 and still refuse to sell as it reaches $15.
As another example, suppose the decision for launching a rocket into deep space depends on 10 critical variables that all must be within a specific range during the launch – temperature, humidity, wind, safety checks, fuel checks, and so on. In this case, the dashboard will give everyone a real-time visual display of the status of the 10 variables and might even serve the role of preventing a launch if any of the variables aren’t green.
Many of these examples are already happening, but the goal here is to highlight how this technology fits into the Intelligence Cycle, to help you develop a spy mindset and shape the way you conduct the business of intelligence.