Weekly Blitz: AI Should Remove Desk Work, Not People: A Real Workflow Test
AGI - Advance, Grow, Innovate with AI
Who’s talking
2 speakers1Transcript135 turns · 50 min
Good morning, Julie.
Good morning, Jason. How are you?
I'm good. How are you? I'm
doing well. Getting ready to sport today.
Oh, yeah, that's right. You're playing in the celebrity baseball game for Bess Bukowski and Joyful
Journey.
Joyful Journey. That's right. We'll make sure there's a link in here. So people want to donate. What does Joyful Journey do?
Adult day services. So kind of mainly for people, elderly, right? But mainly for people to kind of get a little respite. So if they are a caretaker of a parent, they can, you know, talk to Joyful Journey and they usually work with people with a little bit of memory care stuff as well. But they love to have new friends join them and, you know, make that process a little bit easier for everybody.
Well, that's a little near and dear to your heart right now, right? Yeah. You've been dealing with some things with elderly parents, and I'm getting elderly, so, you know. You weren't with me last week.
Yes, I was.
Was it last? Oh, was it a week before that Nirav joined? Okay. But was it last week when I made you the promise to spin up a dashboard and walk through? Yeah. Yes, I have said some words that we cannot say or else YouTube will downgrade us this morning as I was trying to finish that. So it's important to show that AI doesn't always do exactly what you want it to do. And sometimes you have to iterate with it a little bit before you can get the answers you're looking for.
Interesting. Okay.
So shall we go through those things? Do you remember what you told me was like the bottlenecks or the pains in your processes?
Yeah.
The
prospecting, the backend, putting in orders, those kind of things. Yeah.
Right. So I took that right after we got done and I gave it to Cloud or to ChatGPT Codex and had it use some Cloud agents and do some research and generate some best practices. And then I got busy and left. And then this morning, about 20 minutes ago, I was like, oh, damn, did I ever finish that dashboard? So I found the conversation again and I was like, OK, give me here. Here's a picture you can add to the dashboard, create a dashboard. And present it to me. And here's their website. So make it nice and on brand and everything so we can present this this morning. And I'm going to go through the three iterations of dashboards. I want to just kind of show you what it came up with. All right.
So
I'm going to go to my share screen. So it's beautiful. It's on brand.
Got the
picture of you and your bestie there.
Yeah.
Isn't that fun?
Yes, Shelby. Yeah.
Now, I'm not going to tell you what this is talking about. I just want you to look at it and see if it makes any sense to you.
You just got to trust the dang process. Not you. Don't you say that? Maybe.
Maybe.
Move out the credit. What changed because of it? I don't know what that is. I can't see, you know, it's on my phone. So I'm like, oh.
It says a confident dashboard can still mistake correlation for causation. Real comparison measures. What would have happened without marketing? Let me get down here to the ads looked effective until eBay turned them off. Like, I don't, I have no clue what the F it's talking
about. Yeah.
branded keyword ads i'm like what what what are you talking
2Episode notes
EPISODE DESCRIPTION
Jason promised Julie Koehrer an AI workflow that could take prospecting, proposal prep, order entry, and follow-up off her desk. The first two attempts were beautiful, polished, and completely wrong. The third finally found the point: AI should remove desk work, not Julie.
That failure becomes the most useful part of the experiment. Jason and Julie talk through why context matters, where human judgment still belongs, and how much imperfection is acceptable when AI saves hours of repetitive work. Jason also offers Julie a one-month paid ChatGPT experiment so they can test the workflow in public.
The conversation moves into local AI video production with MiniMax H3, the flood of automated solicitation on LinkedIn, and the growing fight over data centers, power, tax incentives, and public accountability. The practical thread holds: use AI to create more room for relationships and judgment, then make companies and public officials answer for the costs pushed onto communities.
ABOUT THE CO-HOST
Julie Koehrer is an on-air personality and branding and marketing consultant with Q106.7 and Woof Boom Radio in Lafayette, Indiana. She brings nearly two decades of radio experience and a practical marketer's skepticism to The Weekly Blitz.
CO-HOST LINKS
Julie Koehrer on LinkedIn: https://www.linkedin.com/in/juliekoehrer