AI for Data Collection is Here!
By Peter Sanderson
June 30, 2023
Originally Published in Quality Magazine
Download the original published article PDF scan
When I was asked to write an article on data collection software for Quality, I had to pause and think about what exactly data collection is without data analysis and actions. Data collection can only be determined by understanding the resulting actions desired after data analysis.
Today we are bombarded by data from the time we wake up until the time we go to sleep, and we are constantly absorbing it, analyzing it, and making decisions about our lives. The data comes to us through our phones, our computers, the television, conversations with colleagues and friends, and based on that data we make choices or decisions.
In business, we collect data using surveys, measurements, and observations, and from online automated sources such as application programming interfaces (APIs). We also collect data based on the analysis of said data that we’ve already collected. This type of analyzed data is generally used to make decisions, such as opportunities for improvement, adjustments to processes, or in severe cases, rework and recall.
There are many different methods for collecting data in an organization, and when searching for data collection software, it becomes clear that there are endless solutions for all industries. No one solution is necessarily better than another. However, specific data collection methods become preferable once an organization has clearly defined the type of data required to support the actions or interventions they want to achieve.
Rather than focusing on existing data collection methods, it is more valuable to look at the future of data collection, data analysis, and resulting actions. The reality is clear—artificial intelligence (AI) is that future.
Artificial intelligence has arrived and will fundamentally change the entire environment of data collection, data analysis, and real-time action. Not only is AI here, but it is already being used and can be applied immediately in many areas.
AI will enable real-time measurement, data collection, analysis, and immediate intervention to prevent out-of-tolerance conditions. As organizations move toward integrating cameras, machines, and intelligent systems that can learn and act, quality management will become fully process-driven, and nonconformances may eventually be eliminated.
To prepare for this future, companies must begin investing in the necessary infrastructure. This includes moving toward modern data architectures such as a data lakehouse.
However, AI is not just a future concept—it already provides practical tools for today. Many organizations struggle with root cause analysis despite using traditional tools such as fishbone diagrams and the five whys. AI now offers a powerful alternative by acting as an expert assistant to help determine root causes of nonconformities.
For example, using AI tools such as chat-based systems, it is possible to input a real corrective action problem and receive immediate feedback and structured analysis.
Consider this scenario:
Problem Description:
Welding rods were found in the welding department without a material lot number during an internal audit. The rods were tagged, removed, and quarantined. Products using these rods were recalled for material analysis.
AI Response Approach:
AI can generate structured root cause analysis using the Five Whys method:
- Why were the welding rods found without a material lot number?
- Why was the lot number not recorded or attached?
- Why was this not detected earlier in the process?
- Why was there no system ensuring traceability?
- Why was that system not implemented or followed?
By working through these questions, organizations can identify root causes and implement corrective actions to prevent recurrence.
Starting today, companies can use AI tools to solve problems using existing data and quality systems. AI can also assist in writing procedures, developing workflows, and recommending what data should be collected for specific processes.
The more significant challenge, however, is planning for full AI integration—where real-time data collection, analysis, and intervention become standard, ensuring consistent processes without deviation.