About Attila Dobai

Who I Am Now

I help organizations navigate AI transformation by treating human wisdom as the multiplier, not the cost. The Human-Centered AI Framework is a structured, five-phase methodology for AI integration built on twenty years of operational excellence practice. It exists because most organizations are responding to the most significant transformation in fifty years by asking the wrong question.

They ask what AI can replace. The better question is what AI can multiply.

Where This Started

I spent eight years as a Meteorological Oceanographic Analyst Forecaster in the United States Marine Corps, finishing my career as a staff meteorologist for HMX-1, providing aviation weather briefings for the pilots who flew the President of the United States. Precision was not a preference in that role. It was a condition of the mission.

What the Marines built into me was a way of thinking. To understand what the weather will do, you start at the planetary scale, the three-cell Hadley structure that drives global circulation. Then you zoom to the synoptic scale, how land, water, and the Coriolis effect shape regional patterns. Then the mesoscale, the continent-level influence of mountain ranges and coastlines. Then the microscale, what that specific ridge does to the marine layer at the airfield at Camp Pendleton, at this altitude, at this hour.

Fail at any scale, and the forecast fails. Succeed at all four simultaneously, and you see what nobody else can see.

The Translation

When I left the Marines in 2002, I spent months pitching companies on a simple idea: the analytical skills I had developed in the military were directly applicable to business. Nobody took the meeting seriously. Eventually, one company gave me a chance as a data entry clerk, temp-to-hire. Budget Truck Rental, a division of Avis Budget Group.

I was processing phone bills. Thousands of landline charges for dealer locations across the country. Most people would have processed the bills. I looked at the patterns.

I found $60,000 a month in unnecessary charges. The root cause was a daisy chain of forwarded lines. When dealers left, their numbers were not disconnected but redirected, then redirected again, accumulating dead forwarding loops across the network. Nobody had mapped the system end-to-end. I did.

I was in my early 40s before I understood that most people do not think this way, holding multiple scales simultaneously, looking for the systemic structure beneath the surface pattern. The multi-scale thinking the Marines had trained into me was not standard equipment.

The Proof

From that data entry role, I rose through Avis Budget Group on the strength of one consistent outcome: solving problems that others had declared impossible. The clearest demonstration came when I was asked to take over the maintenance authorization department for Budget Truck Rental, specifically because it needed to be fixed.

In my first month, I saved a 1,000-truck contract that was on the verge of being lost. Over three years, I cut $40 million in annual maintenance expense from a $100 million budget. I did it without reducing fleet size. Truck uptime increased. No one was involuntarily laid off.

The headcount reduction happened through attrition. When people retired, I evaluated whether the role still needed to exist and what it actually required. I discovered that the role did not require mechanical expertise. It required analytical thinking, attention to detail, and a great attitude. I replaced retiring specialists with people coming out of fast food and similar service roles, who earned more than they had before, developed new skills, and excelled. The new staff moved up. The company saved $40 million. Nobody lost. Everybody won.

The Pattern

What I kept noticing, across projects and across years, was that the same approach kept getting named the same way by the people on the receiving end of it. A supply chain professional, a healthcare operations leader, continuous improvement practitioners across different industries, independently described the methodology using the same phrase: human-centered.

None of them had read the same books. They were describing what they observed.

The Convergence

When AI arrived as a serious force in business operations, I recognized the pattern immediately. The same mistakes, at a faster pace, with higher stakes. Organizations treating AI as a cost-cutting instrument rather than a capability multiplier. Automating functions without asking what the human becomes. Imposing efficiency without designing for it.

I have a phrase for what happens when organizations implement AI on top of processes they do not understand: solutions that metastasize.

The Human-Centered AI Framework is the result of taking what I already knew, from Lean Six Sigma, from twenty years in the field, from the pattern clients kept naming, and building something purpose-built for this moment. Five phases. A structured discipline for deciding what AI should do, what humans must own, and how to design the handoff between them so that both get better over time.

It is not anti-AI. It is pro-human. Those are not the same thing, and the difference matters more now than it ever has.