A few years ago, I was deep into building an RPA function for a large global finance and accounting organization. We were automating high-volume processes — AP, reconciliations, the kind of work that eats operational capacity. And I was sitting in vendor presentations where everything looked clean and fast and inevitable, then going back to my team where the reality was messy, political, and slower than anyone wanted to admit.
The technology worked. But the gap between what was promised in the boardroom and what was possible in operations — that gap was where all the real work lived. Bridging it took more organizational patience, process discipline, and honest conversation than any vendor slide deck would suggest.
Now I’m watching the same cycle play out with AI, at a speed and scale that makes RPA look like a rehearsal.
The promises are bigger. The pressure on finance and accounting leaders to “do something with AI” is more intense. And the people actually responsible for making it work — the directors and managers running operations — are stuck in the same spot we were: trying to separate what’s real from what’s marketing, with limited time and limited tolerance for getting it wrong.
That’s why I started writing here.
The short version of the career
I’ve spent over 30 years in accounting, finance, and operations. I’m a CPA and CMA. My career has moved through roles that shaped how I think about process and technology in ways I didn’t fully appreciate until later.
Early on, as a plant controller at a manufacturing company, I was trained in Juran’s quality improvement methods — cost of quality, Pareto analysis, the idea that improvement follows a disciplined sequence. As a division controller, I learned Deming’s approach to total quality management and started to see process improvement less as a toolkit and more as a management philosophy. Then, as corporate controller at an automotive supplier to Toyota, I got an education in the Toyota Production System. Watching TPS work up close is what made continuous improvement click for me permanently. The balance of customer focus, respect for the people doing the work, and relentless waste reduction — that combination felt right, and it’s been the lens I look through ever since.
That foundation is why, years later, I was the person tapped to lead RPA implementation at scale. Process improvement and technology aren’t separate skills in my experience — they’re the same discipline applied at different layers. RPA taught me what happens when you automate a broken process (you get a faster broken process), what governance looks like when bots are running across time zones, and why the pilot-to-production gap is where most automation programs quietly die.
What I’m doing now
I’m learning AI the way I wish someone had documented RPA — out loud, with the mistakes included.
I’m not positioning myself as an AI expert. I’m a finance and accounting practitioner with a long track record in process improvement and automation who is working through the same questions my audience is: Which tools actually deliver? What does a responsible implementation look like when the technology is probabilistic instead of deterministic? How do you maintain controls and auditability when the system reasons instead of follows a script? Where does the hype end and the operational reality begin?
I’m documenting what I find, what I get wrong, and what I’d tell a colleague over coffee if they asked me where to start. If that sounds useful to you, the best way to follow along is through the newsletter.