“Go AI Everything”. Had a couple of conversations this week that gave me flashbacks...
One IT Director had been told by his management to go “AI Everything”. Another IT manager had been told “find use cases for AI”.
You may be thinking “Wow, what smart, progressive companies! I want to work there!”
On the contrary, I’ve seen this movie before. This top-down “go use X” can lead to a ton of money spent without a lot of results. You’ll get “compliance energy” that goes away when the boss is not looking.
AI is a technology. It’s an approach to solving a particular problem.
Here’s a better approach: get people to understand Lean and the fundamentals of eliminating waste.
Defects, overproduction, waiting, not utilizing talent, excess processing, etc. What you can see on a physical assembly line exists in knowledge work.
Before you “AI everything”, visualize how the work is flowing in your organization. Could be coding customer requests. Could be processing loans. Could be receiving products, servicing them, shipping them back out. Whatever.
Look for the waste and ask:
One IT Director had been told by his management to go “AI Everything”. Another IT manager had been told “find use cases for AI”.
You may be thinking “Wow, what smart, progressive companies! I want to work there!”
On the contrary, I’ve seen this movie before. This top-down “go use X” can lead to a ton of money spent without a lot of results. You’ll get “compliance energy” that goes away when the boss is not looking.
AI is a technology. It’s an approach to solving a particular problem.
Here’s a better approach: get people to understand Lean and the fundamentals of eliminating waste.
Defects, overproduction, waiting, not utilizing talent, excess processing, etc. What you can see on a physical assembly line exists in knowledge work.
Before you “AI everything”, visualize how the work is flowing in your organization. Could be coding customer requests. Could be processing loans. Could be receiving products, servicing them, shipping them back out. Whatever.
Look for the waste and ask:
- How long does this process take?
- Where are the bottlenecks?
- Why does the work product from this step to that step contain defects?
- Why are we always waiting on that team?
- Etc.
From there, focus on the constraint in your process. If we launch new initiatives in QA but they’re never the bottleneck, we’re wasting money. If we speed up loan sales without underwriting, we’ll just increase work in progress and the overall throughput will slow.
(I once worked with an executive, who, upon being hired, immediately announced twenty different quality initiatives at the same time. Employees chuckled. It wasn’t lack of good ideas that kept these from being executed. The constraint was the quantity and time of Sr. Engineers who could lead them.)
Next, what IS the best way to eliminate that constraint? It may be a work flow that has too many approvals. Should we “Use AI” to automate getting those unnecessary approvals? Issue may be staffing. It may be the user interface, employees don’t understand so they work around the system, etc.
From there we can talk technology. You might need a Copilot? You might need a new API integration. You might need a DAP like WalkMe. But without understanding where the waste is, you’re putting the cart before the horse.
What do you think? Am I off base here?