Adaptive Process - Part 1: Human Processing

Lets take an everyday, everybody process to start by way of understanding.

You are going to the shops to buy your weekly groceries. You carry with you a single re-usable bag, your credit card and, because you’re fairly well organised, a shopping list of the things you need to buy.

You also have your assumptions. You only have one bag because you’re only buying for one person, you have your list which is updated based upon things you know you’ve run out of and you know roughly how much you’re going to spend and an awareness of how much you have available to spend.

When you get to the shop, you get to test your current information against Reality, its an experiment as soon as you leave the house. Variables hit you, were you hungry when you went? Maybe a bit more ‘fast’ food is okay. Is the washing powder you need on sale? maybe I’ll buy a bit more. Is the shopping list complete? I need to update it.

So you leave the shop, with all of your items ticked off the list, a few extra items to either offset future costs or satisfy immediate wants. You get home and immediately get feedback on your success or failure. Did you get everything you actually needed? Did you spend more money than you expected? Did you have enough carrying capacity to make it?

If we were going to model this using a business notation it would look a little like this:

adaptiveprocess-01.png

When defining a business process you define tasks, decisions, transactions and of course a beginning and end. These nodes can be further defined using the logic of completion, what action is performed? What are the inputs and outputs? Are there any transformations or procedures involved?

This is also where assigning controls and safety nets can be done. For example, in the ‘Collect Shopping List’ task a safety net to go and check the pantry and fridge for missing items could be included. The ‘Go to Shops’ task could include a web site check for specials and discounts.

These are the checks and balances that at the end of the day a global measurement uses to decide if the process was a success or a failure.

So, how can this process be “adaptive”? Firstly adaptive process can be defined a number of ways:

  • The ability to create an information system that is ‘aware’ of it’s own processing.

  • A non-anticipating process that requires iteration in order to measure success.

  • A method to create and re-arrange known predefined objects, content, business rules and social interaction in order to achieve specified goals.

As can see using the shopping trip example, the processes you do individually can be defined as adaptive. You will review the latest iteration to measure success, you will measure and apply the controls and safety nets, you are the only one with complete awareness of your processing. You will alter the process after each iteration to include new controls, new rules which can improve your outcomes, time taken and positive outcomes gained.

We have developed CERES to take advantage of this naturally occurring phenomenon, to rapidly build, iterate upon and measure success using easily defined objects, content and controls.

Over the next few weeks I’ll show you how adaptive processing can be secured, how it can be scaled to support small and large teams, and how data quality can be assured by using lineage and version control caused by adaptation.