Lynda ArtesaniLegal Accountant and Data Migration Specialist at The Proper Trust

The “Sandwich Approach” to Using AI #legalaccounting #lawfirms #ai #judicial

30 Mar 2026 · 1 min

Watch on YouTube

0:48

In this brief segment, the speaker introduces the “sandwich approach,” a simple framework for responsibly integrating AI into work. The core idea is that AI should never initiate a task nor complete it; instead, humans should start with their own thinking, wrestle with the initial blank‑page challenge, then bring in AI in the middle to refine, augment, and document the work. Finally, the human practitioner must apply their own judgment and verification before finalizing the output. This three‑stage model is presented as a safeguard to preserve human skills, ensure accountability, and maintain quality while still leveraging AI’s strengths.

Chapters

And I talk about this framework that I use, which sounds silly, but it's really simple and in my mind, a really good way to keep a check on myself. I call it the sandwich approach, right? AI should never start it and should never finish it for you, but it's okay to have it in the middle. All right. Right? And to make sure that you are starting with your own thinking, in part because you don't wanna lose those skills. It's important for us to wrestle with that blank moment where we're having to figure out, how do I start this? What do I wanna do here? Really wrestle with that and then use AI, as you say, to refine, to augment, document to help you make it better, and then make sure that you're using your own judgment to verify at the end, right? So AI in the middle and your own human judgment and verification processes on either end, I think, is the safest way to think about it right now.