Prepare for AI: determine the state of your QCs
When considering an investment in Artificial Intelligence (AI), you may want to review the current state of quality checks (QCs) in your organization’s workflows. Some QCs are performed in a timely manner, mitigating the risk of error and negative impact. Some workflows don’t have QCs, but should. Some QCs are skipped, but shouldn’t be. Some are performed too late. Other QCs are so mundane that the mind wanders and one sees what one wants to see, instead of what’s really there. Some QC processes require human thought and analysis. Some may already be automated. Determining the current state of your QC processes before investing in AI is an important preparation step, as AI can be used to improve QCs.
To determine the current state of QCs in workflows, perform the following steps:
Identify all workflows in a department or business unit.
Break down each workflow into a chart showing high-level task descriptions, organized sequentially.
For each high-level task description, include key data points. For example, in the task to onboard a new employee, key data points might include name, address, SSN, and email address.
Identify all QCs occurring during each high-level task and whether the QC is performed by a human or automation. Include formalized QC processes and informal processes (such as an employee performing an unnecessary QC).
Determine pain points for each existing QC. If it’s being skipped, why? If it’s performed too late, why? Is the QC even necessary?
Identify QC gaps in each workflow - a place where a QC should be performed but isn’t. This could be based on the criticality of the data in the workflow. For instance, let’s say I’m responsible for new-hire onboarding. One of my tasks is to enter the new employee’s personal email address in a data field. This email address will be used to deliver the offer letter and other important documents. I must enter the email address manually into our database. Unfortunately, I fat-finger the spelling. Since I don’t QC my own work, and my company doesn’t have a QC process for this workflow, my error causes the following chain of events:
The email containing the important documents is sent to a wrong email address.
The new hire doesn’t respond to the email.
I text the individual and state that the paperwork must be completed and returned ASAP.
The individual responds via text that the email was never received.
I send the email, again. Same problem.
The individual asks me to confirm the email address and provides the correct address.
Upon review of the data I typed manually, I’m still blind to my own error, and send another email to the wrong address.
We are now both quite frustrated.
Only upon the insistence of the new hire, do I finally “see” my mistake and correct it.
The individual receives the email with the documents, completes them, and returns them.
An email address is an important data point for new hire information. Since the data is entered manually, it should be checked. Therefore, the lack of a QC on email addresses would be identified as a GAP on the QC Status document.
7. Create a QC Status document which records all information obtained during steps 1 - 6 above.
When your organization invests in an enterprise AI tool, the QC Status document can then be uploaded to the tool and prompted to generate new methods for performing QCs.
Effective Continuous Improvement begins with clarity on workflows, procedures, and training plans.
This blog was written and edited by a human!
© 2024 Lori K. Barbeau