In the pursuit of finding practical applications for AI, I realized that AI's role in QA can be extremely impactful to reduce error rates and find more efficiencies.
Recently, I came across an issue where there was a mass communication scheduled to be sent at a specific date and time, which contained time-sensitive content; sending too early would have a negative impact, and sending too late would not achieve the desired effect. Thus, the communication was scheduled through the tool, as you normally would. However, due to human error, the scheduled time was incorrect (by days), despite the usual safeguards (I.E naming the communication, checking a box in the planner, etc.) Resulting in the obvious and a whole lot of damage control.
This is just one example of many I can think of.
I had a long discussion with Mrs. Kimberly Sivananthan, MPH, PMP who is an SME and leader in Quality Management Systems and, process implementation in STEM, and realized two things:
Mistakes like this within Life Sciences and Pharma can be much more devastating than some lost sales. Live can be impacted.
To avoid these mistakes, there is reliance on additional human intervention within the processes.
So what is the Role of AI in Quality Management?
AI can be leveraged in a context-aware mode for QA. You already have a glimpse of this with your email tools, where if you write a sentence that states "attached is the XYZ" and you try to hit send without the attachment it warns you.
Role of AI in Quality Management can have a diverse meaning depending on the context it is used. Having the right expertise to understand the use cases and applying the correct AI is crucial to get the expected results.
Take this concept and expand upon it: Have AI process the entire content and understand the context, maybe even understand current news, to then trigger warnings or escalations within the organization.
Some examples I can think of:
A marketing campaign is scheduled to be sent out tomorrow but there is a late-breaking event that makes the campaign's timing ill-fitting. Can we have AI provide a warning and pause until leadership's explicit approval?
What about if there is a campaign scheduled to go out tomorrow, for a sale that starts next week, but the content reads "Shop Today"? Can AI understand that the context is wrong and trigger an action?
Someone has inadvertently attached/entered confidential information into an email that should not be sent outside of your organization. AI can help trigger an explicit approval or warning.
In all of these cases, AI will need to be trained on enterprise data to understand the context better. We have a ways to go, but it won't be long before we can have AI help reduce human error, contextually.