5 Comments
User's avatar
A.E. Lee's avatar

Nice breakdown. My only concern having done Salesforce automations for a long time is that “oh yes this a very straightforward simple process” often turns out to have many edge cases and subtle complexities. Then the boss mandates it gets automated anyhow, and employees complain the automations don’t work since they often have to “fix” the outputs or learn to work with the automation to get the right result. I’ve been recently working in testing and implementing agentic workflows in those cases to handle all the edge cases and nuances. But on the surface, those tasks look like “workflows”. Just a thought.

Expand full comment
Pascal Biese's avatar

Good point! In industry settings (especially in legacy enterprises), it's often hard to judge the complexity of an automation beforehand. I think it's important not to overcommit to a technology and to establish feedback loops as early as possible.

Expand full comment
AI DailyPost's avatar

Neat!

Expand full comment
Johny "One" Purple's avatar

It is happening. I did help build several agents in pas year, now the startups are moving to full agentic ecosystems, several agents working together on jobs to be done:) question is: Will you work for AI or will the AI work for You? Next 3 years will decide https://www.kwisatzh.com/p/architects-vs-workers-will-you-work

Expand full comment
Rajiv kumar Pandey's avatar

Only 🆓

Expand full comment