Week 7 1— Well Oiled Machine
A Tale Of Working in the Bureaucracy
It’s the lazy days of summer as the heat takes over. In Ottawa, we are desperate for rain. I’m looking out at my lawn and all I see is dead grass. I can only imagine how much worse it is for farmers!
So before I launch into a post about my weekly topic, I am going to say that this weeks post is shorter than usual. I’m writing this post later in the day after a long day of meetings and work. If there is a telltale sign of summer, it’s when the grind of day in and day out is getting tougher and the mind starts to wander to time off and a change of pace.
Last week, I talked about risk aversion in the public service. I shared a personal experience from years ago where a successful project was seen as a bad thing because it didn’t follow established process. I further theorized that the success or failure of a project had as much importance as whether or not process was followed. Continuing that line of thought, I speculated that process itself was being used as a cover/shield so individual performance was less important than having a scapegoat in the form of process to lean on/blame.
This week, I want to share how a system which favours established process doesn’t mean a system where things work. In fact, process itself could be leading to failure and any blind adherence to process is risk aversion rather than effective and proven theories of how to get shit done.
Process is established so we have repeatable and scale-able ways of doing things. It also ensures a level of transparency, accountability and consistency to how something will be done which can be especially important if the activity is in support of a legal or regulatory requirement. Process may also help cut costs, processing time or provide any other number of efficiency benefits as it streamlines how things are done. This is not meant to be an academic or streamlined list of the benefits of process but more to paint an overall picture of why process exists.
On the other hand, process is not perfect. Imperfect or inefficient process exists. Process can become gospel where it is blindly followed and rarely questioned. Unless an organization dedicates itself to continuous process improvement, some processes can be unchanged for years even as both internal and external circumstances change. In other cases, process can lead to unintended or negative outcomes. The key point to remember is that process (like many things in life) is neither inherently good or bad and the truth is somewhere in the middle.
As many readers have likely experienced, process can lead to questionable decisions or what seems to be opposite of common sense. So if we’ve established that process is neither good nor bad, then what comes next. Well for one, it means that process does not equal success. It also means that process should not be cover for organizational or individual failure. In fact, following bad process without challenging it is in itself the bigger failure. An organization which stagnates and relies on outdated process from different eras to operate within a modern context is not a healthy organization. Employees who blindly follow process because “it’s how things are done” are as much a problem as the organizations who refuse to perform introspection and challenge themselves to improve.
Process is a double edged sword. It can be used for both good and bad. A blanket statement of process is good or process is bad is a simplification of the issue that isn’t valid. Rather, organizations and individuals should take ownership and autonomy over process. Individuals should feel empowered to challenge process and try new things (within reason… nothing illegal). Organizations should support continuous improvement of their process so stagnation doesn’t set in. These are the conditions of a healthy organization that knows how to get shit done.
AI Demonstrator Projects (Regulatory Metadata, Regulatory Evaluation Platform, Rules as Code and Machine Readable Regulatory Impact Analysis Statement)
Regulatory Metadata: No updates to provide on this project.
Regulatory Evaluation Platform: We are in the midst of evaluating our options for this project. Over the past year, we’ve learned a lot about what it takes to run an AI project successfully. Using this new knowledge, we want to figure out what is the right next step to finish this project.
We are also working on a peer review of this system with some researchers from the University of Ottawa. Based on a deep literature scan, we have a draft outline for a peer review. Next up is a co-design workshop with key stakeholders so we can collaboratively determine the best approach to do a peer review of an AI system. I’m excited to see the results and hope to have something to share soon.
Rules as Code: We have been talking with a few other partners who have interesting use cases using different rulesets. Our hope is that we can work on two Rules as Code projects this year. The one I’ve already written about for several weeks and hopefully one more which i will announce when the time comes.
Machine Readable Regulatory Impact Analysis Statement: We are working on a proposal for this project that will ensure the regulatory impact analysis statement data is made available in a machine readable format.
Week 71 weeknotes are done. I hope everyone has a great week!