I’ve been exploring how AI is starting to move from “assistant mode” into actually executing workflows across different systems, and it’s still not entirely clear where the practical limits are. I recently read this explanation of Large Action Models and how they’re designed to perform structured multi-step actions instead of just generating responses: https://www.trinetix.com/insights/what-are-large-action-models-and-how-do-they-work. It reminded me of a small internal automation setup we had where an AI component could trigger different API calls, but everything still had to be tightly controlled. I’m curious—do you think these systems will ever be trusted to fully execute business workflows end-to-end, or will they always stay in a “suggest and assist” role with humans approving the final steps?
top of page
Mini French Bulldog
Public·233 members
8 Views
bottom of page
I’m not a developer, but I work in a team that uses a lot of automated internal dashboards and workflow tools. From my perspective, the most important thing is predictability. Even if a system is technically more advanced, if people don’t understand why something happened, they lose trust in it pretty quickly. I’ve seen cases where automation actually slowed teams down because they spent more time verifying outputs than they did doing the task manually. So I think adoption depends less on how “smart” the system is and more on whether it feels stable and explainable in day-to-day use.