We asked ChatGPT to fix a production process for a manufacturer.

Can ChatGPT Solve Manufacturing Problem

If you haven’t heard of ChatGTP, you will soon.  It is a new AI that generates remarkably good responses to questions and general prompts from its users.  It isn’t perfect, but it gives reasonably accurate answers to questions about any subject.

There is a lot of buzz in the business community about what ChatGPT can do to help businesses succeed.  There is also some fear that it will replace jobs and make many tasks currently handled by humans obsolete.

Management and Strategy Institute decided to test just how much ChatGTP knew about Six Sigma, and if it could help a manufacturer with a production problem.  The results weren’t stellar, but they still provided enough insight into what the next version of ChatGTP might be capable of doing.

The question we asked:

A table manufacturer has bottlenecks in its production process.  The manufacturer is also receiving complaints about the quality of the tables.  Using Six Sigma, how can this manufacturer fix its problems.

Analyzing the question:

We’ll admit this isn’t the most in-depth question.  We provided ChatGPT with two generic issues, bottlenecks in production and complaints about quality.  We did this purposefully just to see how much information, and how in-depth, the software would go to answer our question.

We’ve noticed that the software tends to produce answers that are roughly 350-500 words for each query.  That tells us that ChatGPT is currently set up to produce answers that summarize the topic and get to the answer quickly.  Here is the answer we received:

ChatGTP’s answer:

Six Sigma is a methodology for process improvement that aims to reduce defects and variability in a process.  It can be used to identify and eliminate bottlenecks and improve the quality of the product.

One way the manufacturer could use Six Sigma to address the bottlenecks in its production process is by using the DMAIC (Define, Measure, Analyze, Improve, Control) methodology, which is a structured approach for problem-solving.

  1. Define: The first step is to clearly define the problem and establish specific goals for improvement. In this case, the problem is bottlenecks in production and complaints about the quality of the tables.
  2. Measure: The next step is to collect data and measure the current performance of the process. This includes identifying the key performance indicators (KPIs) that are most relevant to the problem, such as production throughput, cycle time, and rejection rates.
  3. Analyze: The data collected is then analyzed to identify the root causes of the bottlenecks and the quality problems. This might involve creating process flow diagrams and using tools like cause-and-effect diagrams to identify the key drivers of the problems.
  4. Improve: Once the root causes have been identified, the next step is to develop and implement solutions to address them. This might include changes to the process, such as implementing new equipment or procedures, or changes to the design of the tables.
  5. Control: The final step is to establish a system of controls to monitor and maintain the improvements made. This might include setting up quality control checks at key points in the process and collecting data on the performance of the process over time to ensure that the improvements are sustained.

 

As for the Quality complaints, one way to identify the problem is to perform a failure mode and effects analysis (FMEA), this will help to identify the potential failure modes in the product design, process, or assembly and the potential effects of these failures, allowing the manufacturer to take proactive measures to prevent them.

In addition, Six Sigma methodologies such as Statistical Process Control (SPC) and Design of Experiment (DOE) could be used to monitor, control and improve the quality of the tables.

It’s important to keep in mind that Six Sigma is a process and data-driven approach that helps to identify, diagnose and solve problems.  A Six Sigma project team can help to implement the process by involving the relevant team members and utilizing their knowledge and skills in the problem-solving process.

Analyzing ChatGTP's answer about Six Sigma process improvement:

The Good:

It’s clear that ChatGTP can produce a coherent summary of Six Sigma and give general guidance to solving the problems we presented.  It broke down its answer to both questions separately and provided advice that aligns with the recommendation a Six Sigma Black Belt would give to resolve the manufacturer’s problems.

The Bad:

We can say a lot about the answer that ChatGPT gave to our question, but we can summarize it in one word – lackluster.  The answer ChatGTP provided is essentially a summary of the DMAIC process, with some quality improvement processes sprinkled in.  The answer has no depth, and is something we could have found with a simple Google search.

For example, when discussing customer complaints, the answer mentions SPC and DOE, but doesn’t explain how to use them.  This would be of little use to a Six Sigma Champion tasked with improving the manufacturing process.  In fairness, it would be a good way for a Six Sigma Black Belt to generate some initial project ideas while developing the project charter.  At this point it’s unknown if the software uses a Six Sigma Body of Knowledge to generate ideas.

Things it missed:

Voice of the customer – There is no mention of VoC. That’s particularly odd, considering we mentioned the company was receiving quality complaints.  Since the customer measures quality, you can’t improve a product when you don’t know how it is perceived.

Standard Deviation – To resolve a bottleneck, one would assume that measurements would be required.  The answer doesn’t mention collecting data points to find the normal distribution.

What tools to use – No mention of Run Charts, Pareto, Histogram, Analysis of Variance, or any other tool required to accomplish the improvements.

Conclusion:

One of the nice things about ChatGTP is that it can expand on its answers.  We could have continued to ask follow-up questions and dug deeper into the problem.  Whether ChatGPT will someday replace Six Sigma Black Belts and statistically relevant improvement projects is yet to be seen, but in the short term, it’s safe to say that people will be leading the process improvement journey.

Picture of Rei Takako

Rei Takako

BS, Business Management
Chicago, IL