Aug 5, 2026 · 46 min · 11 segments
Most jobsite software wasn't built to capture what AI actually needs. The fields, timestamps, and check-ins that platforms record today were designed for tracking tasks, not for training models or…
Alex SchnellGuest
Aaron SmithHost
Uh, by the end of it, uh, you will be able to audit yourself this week whether your operation captures the data that AI actually needs.
Uh, the good news is, is all of it is fixable, and none of it requires a prescriptive system or solution.
Uh, there are lots of ways to solve this problem, but it does require your system, whatever that may be, to be capable of holding and capturing a certain structure, and that, that structure is really what we're going to dive into today.
Um, to make this really concrete, uh, Mariel's gonna be putting up a, a poll here, and I wanna walk you guys through a really simple example, uh, where we have a task that finished late.
Um, and a really simple question for all of you guys is, can your system tell the difference between these three distinct things? Can it tell which one of these three things actually happened? The first option is the trade started late, but executed on pace.
Versus they started on time and the work ran long, versus option C, they started on time, finished on time, but the task looked done and then somebody had to come back, so the go-back work actually delayed the true completion.
So the question for the group is, is can your systems tell them apart cleanly, partially, not at all?
And Alec, just, uh, if people go into the chat and you scroll up, you'll see View Poll.
Just click on View Poll and it'll give you the yes, cleanly, partially no, or not sure, um, for our poll today.
Uh, by the end of it, uh, you will be able to audit yourself this week whether your operation captures the data that AI actually needs.
Uh, the good news is, is all of it is fixable, and none of it requires a prescriptive system or solution.
Uh, there are lots of ways to solve this problem, but it does require your system, whatever that may be, to be capable of holding and capturing a certain structure, and that, that structure is really what we're going to dive into today.
Um, to make this really concrete, uh, Mariel's gonna be putting up a, a poll here, and I wanna walk you guys through a really simple example, uh, where we have a task that finished late.
Um, and a really simple question for all of you guys is, can your system tell the difference between these three distinct things? Can it tell which one of these three things actually happened? The first option is the trade started late, but executed on pace.
Versus they started on time and the work ran long, versus option C, they started on time, finished on time, but the task looked done and then somebody had to come back, so the go-back work actually delayed the true completion.
So the question for the group is, is can your systems tell them apart cleanly, partially, not at all?
And Alec, just, uh, if people go into the chat and you scroll up, you'll see View Poll.
Just click on View Poll and it'll give you the yes, cleanly, partially no, or not sure, um, for our poll today.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
