Q and A with DLS Alumnus Michael Carsley, ‘88, President of Carsley Solutions

Carsley majored in Computer Science at Michigan State University, with minors in Databases and Artificial Intelligence (AI) where he used LISP, one of the first languages designed for AI and still in use today. 

Carsley founded his company in 1996, and today Carsley Solutions offers numerous AI applications for its clients. 

Where do people encounter AI in their daily lives without realizing it?
AI is already woven into everyday life. When your phone suggests a text reply, when Netflix recommends a show, when Google Maps reroutes you, when your bank flags fraud — that’s all AI quietly working behind the scenes.

When you say ‘Hey, Siri, call John Smith,’ Siri uses natural language parsing to understand your intent, recognizing that you want to make a call and identifying the name from your contacts. AI apps like this one save time by turning speech into action. 

What AI project are you most proud of and why?
At Carsley, I’m most proud of the AI work that helps manufacturers make better decisions — predicting downtime, improving scheduling, and eliminating manual work. It’s not flashy, but it’s real value. It saves time, money, and headaches.

What is one thing you wish more people understood about AI?
AI doesn’t think, it predicts. It doesn’t understand the world like humans do. That’s why human judgment is still essential.

If you could clear up one misconception about AI, that would be?
That AI is going to replace everyone. It’s not. AI replaces tasks, not people. The people who learn to use AI will replace the people who don’t.

What skills should students or workers develop to stay relevant?
I’d focus on three: Critical thinking, problem‑solving, and communication. AI can automate tasks, but it can’t replace good judgment, or the ability to explain your thinking.
 
What is one use of AI that excites you - and one that concerns you?     What excites me is AI that eliminates repetitive work so people can focus on higher‑value thinking. That’s where productivity explodes. 

For example, with the Carsley Solutions Optic AI tool we built, we take mountains of documents — vendor invoices and customer purchase orders— and instead of someone manually keying them in or chasing down errors, the AI reads them, extracts the data, validates it, and posts it directly into the company’s ERP system.

Enterprise Resource Planning (ERP) software connects the core of a manufacturing business — production, inventory, purchasing, accounting, and supply‑chain operations — into one integrated system. By automating the flow of accurate data into that system, Optic AI helps manufacturers eliminate manual entry, reduce errors, and keep operations synchronized from the shop floor to finance.

What used to take hours now takes minutes. People stop typing and start improving processes.

It’s a perfect example of AI augmenting people, not replacing them — turning repetitive paperwork into real‑time manufacturing intelligence.

And an AI usage that concerns you? AI used without oversight. If you don’t understand the data behind it, you can get biased or flat‑out wrong results. Too much dependence and not enough self‑thinking is dangerous.

AI is a great research tool, but you still must understand the business rules, the use cases, and the customer’s real problem. You can’t take AI output as gospel truth — it’s not always right.

With our “Optic AI” tool for example, we’re very intentional about this. The AI does the heavy lifting, but humans still validate exceptions, apply judgment, and make the final call. That balance is critical.
 
What are some of the biggest ethical concerns in AI right now?
I’d put three at the top: bias, transparency, and accountability. If the data is biased, the AI will be biased. People deserve to know when AI is being used. And someone must be responsible when AI gets it wrong.
 
Who should be responsible when AI makes a mistake?
The organization using the AI. AI is a tool. If a tool is used incorrectly, responsibility sits with the humans who deployed it.
 
What AI will realistically change in the next 5–10 years?
AI will become part of every job — the same way spreadsheets did. Not replacing people, but amplifying them. You’ll see AI copilots in every software tool, smarter manufacturing, faster decision‑making, and personalized learning.
 
What are people not paying enough attention to?
Data quality. Everyone talks about AI models, but the real power — and the real risk — is in the data. Bad data equals bad decisions. People should be asking: Where did the data come from? If you don’t understand the data, you don’t understand the AI.