When you automate your decisions and execution, you may lose the “why” behind it.
I worked with a manufacturing company where one employee was responsible for scheduling. This person knew which job needed to run first, which machine to use, when something could wait, and when the schedule needed to change.
Then that person left.
The company still had the schedules. What they didn’t have was the reasoning behind many of the decisions.
Why did this job run before that one? Why was a particular machine used? What information was being considered when priorities changed? What exceptions did this employee know about simply because they had been doing the job for years?
The daily digest for entrepreneurs and business leaders
An Inc.com Featured Presentation
I think about that experience now when I
I think about that experience now when I see companies giving AI a bigger role in operational decisions. When the human or AI’s reasoning behind an important decision isn’t captured, the business can lose the ability to understand why that decision was made.
As AI starts recommending, making, and even executing decisions, standard operating procedures need to change with it.
What exactly is AI deciding?
Let’s take that same scheduling example. Suppose the company starts using AI to analyze orders, production capacity, staffing, deadlines, machine availability, and other information to recommend a production schedule.
That can save a tremendous amount of time
That can save a tremendous amount of time. But if AI is now part of the process, its role and decision-making authority need to be documented too.
What information is AI supposed to consider? What rules is it following? Is it recommending the schedule, or is it actually making the scheduling decision? What happens when a rush order comes in? What if a machine goes down? Who can override the AI, and under what circumstances?
It is no longer enough for the process to say, “Use the AI tool to create the production schedule.”
That tells the employee what to do
That tells the employee what to do. It doesn’t explain how the decision is being made.
Document the decision, not just the task
To understand the decision behind the processes during documentation, you want to address and capture the employee’s answers to these questions:
Why did you choose that option? What happens if this changes? When do you make an exception? Who needs to approve this? What information do you look at before you decide?
That’s when you start uncovering the knowledge that is running the business.
Those questions become even more important when AI
Those questions become even more important when AI is involved, because now AI may be participating with its judgment.
When AI becomes part of a workflow, its decision rights need to be defined.
I look at this in four levels.
1. AI can inform. It provides information that helps an employee make a decision.
2. AI can recommend. It analyzes the information and suggests what should happen.
3.AI can decide. It determines what should happen based on the information and rules it has been given.
4. AI can execute. It takes the action without waiting for someone to approve it.
You should know which level of authority you have given AI.
Your SOP needs to define the boundaries
Take a process in your company where AI is already being used and look at the procedure.
What information does AI receive
What information does AI receive? What is it allowed to do with that information? What decisions can it make? When does a human need to review or approve the decision?
Then look at the exceptions.
Who can override the AI? When does something need to be escalated? And ultimately, who owns the outcome?
That last question is extremely important. If AI makes or influences a decision that costs the company money, creates an operational problem, or negatively affects a customer, saying, “The AI made the decision,” doesn’t resolve the issue.
Someone still owns that process, and accountability cannot be automated away.
Here’s an exercise I would recommend for your next leadership meeting.
Ask each department leader: Which decisions in your
Ask each department leader: Which decisions in your department are currently being made or influenced by AI?
Don’t stop with a list of AI tools. Find out where AI is actually entering your workflows and influencing what happens next.
Then ask: Is AI informing, recommending, deciding, or executing? And where is that documented?
You may discover that AI has become part of your operations without anyone stopping to update the procedures that govern those operations.
I am a big believer in AI
I am a big believer in AI. I use it and my team uses it. I think the opportunities for businesses are tremendous.
One reason to document processes is to make sure the knowledge of how a company operates doesn’t disappear when someone leaves. As AI is given more responsibility within business processes, it is just as important to document the reasoning, rules, and boundaries behind the decisions AI is allowed to make.
So the next time AI recommends or makes an important decision inside your company, ask: Do you know why your AI made that decision?
Get 1 Smart Business Storydelivered straight to your inbox when you subscribe to Inc.’s free daily newsletter.
Source: www.inc.com
