Human in the Loop: Two Ex-Microsoft Builders on AI
Sep 23, 2026 · 33 min · 8 segments
Your shipping carrier sends a bill. Your invoicing system sends invoices. Nobody has ever compared the two, because the systems were never connected and the reconciliation was a spreadsheet import…
Yeah, I mean, there have been numerous instances where we've uncovered things using AI tools to just ingest those data sets.
And so even if the reporting is bad, we can kind of build what we need on top of that using an MCP or even if it's just direct Postgres access or something like that.
So we found errors on the cost side where we weren't charging or billing and invoicing for things that we should have been, that it's picked up on because it was reconciling.
And those were not connected together in any way It was mostly manual, you had spreadsheet imports and stuff like that.
And then we extrapolated that out over about a year and looked at it and we found 600 something thousand dollars of un-invoiced charges that we paid for as a company that we never paid.
passed on to our our clients and so there are things like that and then it's also caught some um some overcharges that allowed us to go back to our shipping carriers and file claims to get that money credited back um and then the other side of it you know on the revenue side we've we we've been using some mcp tools that we've created with our e-commerce platforms to identify products that have recent movement activity, but are kind of hovering around a low threshold number that the CSR, the sales rep may not have been paying attention to.
And so we built proactive reporting that triggers draft emails to the client on behalf of the CSR that says, hey, these flyers are burning through your inventory pretty quickly.
Do you want to re-up these? And so it's allowed us to bring in some revenue in ways that we may not have been catching before.
And when you describe that, looking back, it's so obvious and it's so like and I bet it was relatively easy in the grand scheme of us, you know, talking in all caps to our to our cloud agents in the grand scheme of that.
either not having that thought process yet not having someone maybe like who has that idea a way to lift that voice up um but it's it's brilliant and it's so simple and it's such an easy way to like to build relationships to to to do upsells i love it man it's awesome
And if you think about it, like the uncovering data errors, which resulted in missing revenue or something, you found it once with AI, that becomes a proactive process you run.
Now it can be scheduled or you build something to prevent that error from happening.
Yeah, I mean, there have been numerous instances where we've uncovered things using AI tools to just ingest those data sets.
And so even if the reporting is bad, we can kind of build what we need on top of that using an MCP or even if it's just direct Postgres access or something like that.
So we found errors on the cost side where we weren't charging or billing and invoicing for things that we should have been, that it's picked up on because it was reconciling.
And those were not connected together in any way It was mostly manual, you had spreadsheet imports and stuff like that.
And then we extrapolated that out over about a year and looked at it and we found 600 something thousand dollars of un-invoiced charges that we paid for as a company that we never paid.
passed on to our our clients and so there are things like that and then it's also caught some um some overcharges that allowed us to go back to our shipping carriers and file claims to get that money credited back um and then the other side of it you know on the revenue side we've we we've been using some mcp tools that we've created with our e-commerce platforms to identify products that have recent movement activity, but are kind of hovering around a low threshold number that the CSR, the sales rep may not have been paying attention to.
And so we built proactive reporting that triggers draft emails to the client on behalf of the CSR that says, hey, these flyers are burning through your inventory pretty quickly.
Do you want to re-up these? And so it's allowed us to bring in some revenue in ways that we may not have been catching before.
And when you describe that, looking back, it's so obvious and it's so like and I bet it was relatively easy in the grand scheme of us, you know, talking in all caps to our to our cloud agents in the grand scheme of that.
either not having that thought process yet not having someone maybe like who has that idea a way to lift that voice up um but it's it's brilliant and it's so simple and it's such an easy way to like to build relationships to to to do upsells i love it man it's awesome
And if you think about it, like the uncovering data errors, which resulted in missing revenue or something, you found it once with AI, that becomes a proactive process you run.
Now it can be scheduled or you build something to prevent that error from happening.
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.
