I want to share my thoughts on the importance of quality assurance in information management based on my extensive experience in the field, where I found it has been given short shrift in many organizations – particularly when implementing newer technologies. I'll discuss how even small issues can have significant ripple effects on an entire system and why thorough testing is essential to maintain data integrity.
I often compare the role of quality assurance in information management to the butterfly effect, where one small thing can have a major ripple effect on how everything else functions. This is why I always emphasize the importance of quality control and testing. In my experience, some people in Agile environments focus on completing individual steps without going back to test whether one step has impacted another. This oversight can lead to a muddy and problematic situation.
I've implemented systems and worked with companies that prioritize testing for decades. In fact, when I was at Bell Labs, I was in charge of several testing initiatives. I became known as the person who would walk into a room and elicit moans from everyone because they knew I likely found something in the test that didn't work with their program. However, I always reassured them not to panic immediately, as the issue could be due to bad data fed into their program rather than a fault in their code. We would then embark on a detective-like journey, tracing back the breadcrumbs to identify the root cause of the problem.
I believe that the ability to identify problems is a desirable skill set and a gift for an information leader or practitioner. Organizations need people who can raise red flags when they spot issues like corrupt data sets or missing key fields that could hinder the effectiveness of AI engines or any new system initiative. Without such individuals, companies risk facing disasters that could have been avoided if someone had spoken up about potential problems.
In my current role as a consultant, I continue to emphasize the importance of thorough testing. I recently finished an engagement where I went through the testing phase and identified issues, such as incorrect dates or unexpected behavior when randomly testing the system. I would bring these concerns to the team, asking what had changed and why something that used to work was no longer functioning properly. Although I wasn't necessarily the one diving into the software code, I would investigate the process to pinpoint where the problem likely originated. This persistence in quality assurance is crucial, especially when multiple people are working on different aspects of a project.
Through my experiences, particularly at the Labs where we worked on a huge system with multiple programs, I learned the importance of data hygiene. We developed rules for the database based on the values and their expected locations, which helped us determine when things were correct or incorrect. The integrity of AI systems and other data-driven applications depends on the integrity of the data they consume, including data generated from documents. Quality control is vital when using OCR systems and other digitization methods to ensure the accuracy of the resulting data.
In conclusion, quality assurance plays a critical role in information management. My experiences have taught me that even small issues can have significant consequences for an entire system. By being persistent in testing, identifying problems early, and maintaining data hygiene, we can ensure the integrity of our information management systems and prevent potential disasters. As information management professionals, it's our responsibility to be the guardians of data quality and to speak up when we spot potential issues.
This blog post is based on an original AIIM OnAir podcast recorded on March 5, 2024. When recording podcasts, AIIM uses AI-enabled transcription in Zoom. We then use that transcription as part of a prompt with Claude Pro, Anthropic’s AI assistant. AIIM staff (aka humans) then edit the output from Claude for accuracy, completeness, and tone. In this way, we use AI to increase the accessibility of our podcast and extend the value of great content.