A mechanic in an auto shop can run a computer diagnostics program on your vehicle to identify what’s wrong, but he can’t do the actual physical trouble-shooting himself. A customer service rep can take you through a scripted checklist for trouble-shooting a problem with your air conditioner, but after you’ve exhausted this checklist, you’re both stumped. Meanwhile in IT, the crackerjack no-code developer writes an app and deploys it at record speed, but he’s at a loss when the app uses more resources than it should and needs to be tuned.
All are examples of how fundamental business processes, and the IT behind them, have become so abstracted away from the actual organic process of doing something that the employees who are charged with performing these tasks simply cannot do them if the predetermined recipe for task performance fails.
In a visit I had with a materials engineer in the semiconductor industry, one manager confided that he was deeply concerned that a new generation of material engineers lacked the ability to “develop workarounds” when a particular metal needed for manufacture was in short supply.
“In my day, we did this,” he said. “But the new engineers just aren’t trained to go beyond the recipes and work scripts that have been defined for them.”
How did we get here?
The streamlining of business processes, highly beneficial in speeding times to market, likely had much to do with it.
There is that flip side of the coin. What do you do when you can’t get answers to an elusive problem — because your workforce has become so abstracted away from the problem — with technology and automation that they really can’t critically think beyond what their tools tell them? As more employees are trained to push buttons and work with automation, often with no substantive knowledge about the “bare metal” processes that these buttons and automation are running, these knowledge gaps can become real problems for companies.
How to make automation work for your company
What can be done so that companies can maintain and grow their internal knowledge bases and work with the advantages that workflow automation gives them?
Design business processes that can handle exceptions
Vendors of AI and automation say these technologies do mundane work so employees can do important work. That’s fine, but what happens if a process exception arises? Will the employee have the skills to address it?
SEE: Artificial Intelligence Ethics Policy (TechRepublic Premium)
One step IT and business users can take in system and process design is to design and test for exceptions as well as standard processing. Are there process steps in place for exception handling? Are there employee knowledge gaps that exist that could prevent exceptions from being handled? What is the project game plan for developing this knowledge for exception handling?
Reimagine the customer and product support process
For years, companies have designed and utilized a three-tier customer and product support system that consists of tier one employees who had little knowledge of the product, but who could answer the phone or begin a chat. Tier two of this process consisted of subject matter experts who could intervene if the initial representatives couldn’t resolve the issue. At tier three, a highly knowledgeable product specialist was brought in. Over time, both AI and automation were introduced to support this process flow.
Let’s imagine that a water system manufacturer redesigns the process by reducing customer problem escalation from a three-tier to a two-tier model. At tier one, a somewhat knowledgeable person who is almost an SME takes the initial call. This resolves roughly 90% of the issues. If issue resolution still proves elusive, the customer is forwarded to a highly skilled product specialist.
IT then works with customer support to redesign the workflow and its systems. The result is an uptick in customer satisfaction. More calls requiring issue resolution are resolved in the first call.
In this example, it’s questionable whether the company boosted its overall internal knowledge base — but the company did relocate this knowledge base so its customers had more immediate access to it.
Keep your old hands on deck — and let them teach!
IT has moved to more process automation and automated code generators so that both IT developers and citizen developers can save time. Unfortunately, this doesn’t change the fact that 70-80% of global business transactions and billions of lines of code are written in COBOL, and that the average U.S. COBOL programmer is in their fifties and contemplating retirement.
COBOL is not a no-code or low-code report generator that does everything for you. It requires significant knowledge of systems and subroutines, so it’s incumbent on companies to recognize their investments in COBOL and the need to sustain these investments.
Central to this is affecting a knowledge transfer of COBOL skills from older employees to newer IT workers.
Can AI compensate for lack of employee know-how?
AI and automation are critical elements that companies must insert into business processes if they are to stay competitive. At the same time, companies need strong internal knowledge bases and knowledgeable employees.
There is no reason to think that companies can’t have the best of both worlds if they plan for automation and conduct gap analyses to see where they might be lacking critical skills.
Interested in learning about Cobol? Check out The Complete COBOL Course From Beginner To Expert.