From Peter Drucker to Elon Musk – A Timeless Lesson for AI

Ready. Fire. Aim.

I’m sure we’ve all heard this before, and that most everyone understands the metaphor of shooting before aiming. Not super smart.

But as businesses quickly get excited, and motivated, and pressured to use AI to improve and optimize their processes, all I’m saying is – duck!!

So, let’s unpack this a bit so no one gets hurt.

AI is Exciting, but Start at Zero – Ask: “Do We Even Need to Do This??”

Before you insert AI or make any improvements to the process in question, first ask if the process is even needed. To quote Elon Musk:

“People have mental straitjackets on. They will work on optimizing the thing that should simply not exist”

And if you think this just a new belief, here is a quote from Peter Drucker, one of the foremost management and improvement gurus of the 20^th century:

“There is nothing so useless as doing efficiently that which should not be done at all.”

Unleash the Power of Optimization with Process Mapping - The Unsexy Secret Weapon

I know it’s unsexy, but cross-functional process mapping is the next step.

Follow the simple “SIPOC” process mapping model to document Suppliers, Inputs (and their requirements), Process Steps, Outputs (and their requirements), and Customers. Start high level and then document the sub-processes below them. Also, remember that when deep-diving your processes, document them cross-functionally. Customers and workflows cross functional verticals, and so must we.

Documenting the processes in this manner will quickly highlight bad or missing requirements, steps that should not be there, missing steps and standards, and an initial view into what can be optimized.

I can’t emphasize this process mapping step enough. Elon Musk famously talks about his five steps for making process improvements, which is very powerful, but the actions he professes can’t be taken until we first get our hands dirty with understanding the processes in question. Here is his five-step framework:

·      Question the requirements of your process

·      Remove unnecessary process steps

·      Simplify and optimize what is left

·      Accelerate cycle time – do it faster

·      Automate

Take note that “automate” is intentionally at the end. This is where AI will likely be leveraged the most.

There is also much talk about AI automating jobs, but we need to think about it differently. AI automates tasks, which could be 10%, 20%, 70%, etc. of a person’s job. These tasks surface during detailed cross-functional process and sub-process mapping. You can’t optimize or automate what you don’t see.

For the people in these processes, the real question then becomes: What new, critical work could they take on that wasn’t being done before? Efficiencies from AI aren’t all just “cost cut” benefits, they are also benefits to self-fund new and better value-added work.

In a recent AI class led by Pascal Bornet, “Agentic Artificial Intelligence: Harnessing AI Agents to Reinvent Business, Work, and Life,” he presented many great data points but two stood out for validating this cross-functional process mapping approach when implementing AI.

1. Documented robust processes are twice as likely to succeed in their AI efforts.

2. Cross-functional AI efforts are 2-3x more likely to succeed than in isolated functions.

It’s exciting to think about what AI can do for us, but take the time to question and then document your processes first. It may seem like the slower path, but that’s just a perception in the moment. Slow is smooth, smooth is fast. Map first, then AI — and you’ll go further, faster.