With DevOps, you can prepare your organization to be more productive and profitable by automating the right processes at the right time with the optimal balance of machine and human interaction.
In an increasingly competitive marketplace that demands more value faster, the word “automation” has become synonymous with “efficiency.” Organizations are scrambling to automate business processes with the hopes of quickly streamlining operations and gaining a competitive edge. After all, when automation works well, it’s a masterpiece of coordination and execution. It removes human error as it swiftly automates processes straight to production. Profits soar, investors are happy, and everybody wins. In reality, automation brings speed to processes – and speed is often mistaken for efficiency. When we automate processes that aren’t ready for automation, we can accelerate and intensify their flaws to produce faster, more spectacular failures. Let’s consider luxury automaker Tesla’s recent automation blunder. CEO Elon Musk announced his plan to fully automate Tesla’s factory assembly line in 2018. By the end of the first quarter, he was forced to admit that excessive automation had actually slowed down production. The over-reliance on robots on the assembly line had created enough chokepoints in the manufacturing process that Tesla missed production goals and the company’s stock took a first quarter nose dive. Mr. Musk learned the hard way that automating isn’t the same thing as becoming more efficient. In fact, it can have the exact opposite effect. Tesla’s assembly line breakdown resulted from over-automating a complex process that required human interaction at critical junctures, as well as tactile precision with some of the finer assembly tasks such as placing fiberglass on top of batteries. Rather than improving efficiency on the line, robots had slowed it down because they were unable to self-correct when unforeseen issues arose. As Musk himself remarked, “Humans are underrated.” Grafting is another practice which can cause the best laid plans to go awry. When we simply overlay existing technology and processes with new technology, it can create serious problems. General Motors’ (GM) infamous automation experiment provides a classic example. In the 1980s, GM went out on a limb to compete with Toyota, which was quickly eroding GM’s market share. They invested in spray painting robots to save time and personnel hours on the paint line. Unfortunately, the robots and vehicles fell out of sync when the computer-guided dolly veered slightly off course. The robots began to misfire, mostly spray painting each other. This colorful fiasco resulted in an expensive production delay since the affected vehicles had to be sanded down and repainted at another facility.
Assessing IT Delivery Process Readiness
Before investing in automation, it is imperative for organizations to evaluate their internal processes to determine IT delivery process readiness. We need to know which processes are ready for automation, which require re-engineering before automation should be considered, and which processes should remain manual indefinitely. There are three main factors to consider when assessing IT delivery process readiness:- Complexity: First, we must consider the complexity of the process. Complex processes can be brittle, and thus, costs and risks associated with every potential break must be factored into the decision. When an automated process creates broken builds faster, it begins to consume more resources than its manual counterpart. The more complex the process, the greater risk of breaking – which generates the need for triaging or fixing the broken builds.
- Process Maturity: Second, we must determine if the manual process has matured to the point that automation can significantly improve it. Is the effort to automate worth the cost and effort required to do it right? Is it worth the time and cost of re-engineering the process to prime it for automation? And should the automation be done right now or deferred to the future?
- Human Interaction: Finally, we must determine which aspects of a process require (or may require) human interaction. Despite constant advances in artificial intelligence and machine learning, some processes have stages or junctures that necessitate human reasoning and tactile precision. Simply put, no robot can do what the human mind and human hands can do. This can range from complex decision making to grasping fluffy fiberglass.
Sound Principles for Process Change Through Automation:
- Introduce automation slowly and start with the bottlenecks. Debug these processes before adding additional automation.
- Prioritize processes to be automated. Focus on smooth running processes first and then move on to processes which may require reengineering prior to automaton. This concept applies to CI/CD builds, test automation, and even Robot Process Automation.
- Refrain from juggling too much change at once. Combining new workers with new technology can result in disaster.
- Avoid grafting new technology onto old processes. Before automating a process, restructure the work so that the process runs smoothly with fewer people required.
- Actively encourage adoption. Major process changes are not adopted throughout an organization instantaneously. Organizations often face obstacles to improved productivity, such as management, cultural barriers, etc. Change requires humility, dedication, and a willingness to change. Change agents must encourage organization-wide adoption by engaging and empowering employees.
- Do not expect instantaneous success. While automation may accelerate some processes, it will usually take time for the organization to realize the benefits of the new process.
