Leverage AI, Don't Fall for the Hype

Humane Work and Process in a Focused AI Team

Let’s build with this stuff. Photo by Jaqueline Fritz

An Introductory Rant

The hype race is run by people who have honed their skills on fear, exaggeration, and false promises. Whether you are talking about eating dogs in Indiana or AI taking all of our jobs by September, there are people who profit from scaring people with information that provides a stable of false devils and demons.

AI has no shortage of charlatans, promising business leaders that it can replace people (it can’t) and making people think they can go and get a degree in something that has existed for 18 months. AI is a set of calculations that right now excels in one thing and one thing only: creating output that is capable of barely passing the laugh test. AI generated output is routinely wrong.

So, we pull out George Box yet again, “All Models are Wrong, some are Useful.”

AI is a model. AI is incredibly and predictably wrong, and incredibly and predictably useful. You just have figure out how to use it. AI is not superman, but it will your kryptonite.

Before you set out to fire thousands of your employees to make way for AI, give those same employees access to AI tools and tell them, “Map your flow of work, find out what’s missing, and create AI agents aimed at removing those bottlenecks”

How can we use AI right now?

If we’d rather not lose millions of years in corporate experience and replace it with alpha-level technology, we need to think seriously about how our work actually flows and where AI can really help.

This is a challenge as almost no one has mapped out what their work requires, how they currently work, and where real problems actually reside. We don’t have an idea of our current state, yet we are ready to lay off thousands of people in a bet that AI can do better?

The intersection of artificial intelligence and human work cannot be about “replacement”. Not only is the technology not there, the loss of institutional knowledge will always outweigh any efficiency gains you are envisioning.

In other words, you need to value your existing systems first. Know your current state and improve it.

Value Streams Will Always Exist

The flow of work we have every day is largely misunderstood. Our interactions are usually governed by assumptions…who you assume does that, what I assume you do, what we assume the customer wants. Value stream mapping is getting the people who actually do the work in a room (might be a virtual room) and we map out the work, the mistakes, the breakdowns, the interactions, the collaborations, the frustrations, the joys … everything about the work. Then we figure out how to make it better.

I’ve been doing these for over 20 years and I’ve never seen one not result in surprise, emotion, and rapid positive change. I’ve never seen a team not be surprised about something in their work.

So consider that. How can you deploy an AI into a value stream that you do not know how it operates? The AI will be under informed, the workflows will be tainted, and you will never know because you fired the people who would be able to see what wasn’t right about the outcome.

Vastly superior, would be to take the work that is being done, map it out with the team, look at where the team is missing information, processing, or resources and then use AI to help alleviate those bottlenecks. Don’t replace people with AI, augment their workflows with it.

Be smart, be humane, ignore the hype, but take full advantage of the new tool.

Want more? Come see us at Modus Institute and sign up for the VSM class…or drop me a line on LinkedIn and we’ll talk about doing a value stream mapping exercise for AI enhanced teams together.