AI Meat Proxies and Workslop

Someone on your team ran a task through AI and forwarded the output without reviewing it. That person was a Meat Proxy and just generated Workslop. Yuck!

Alane Boyd and Micah Johnson break down these two terms, which sound like they should be in food manufacturing, and describe exactly what can go wrong inside AI-forward companies.

In this episode, you’ll learn:

• How to tell a meat proxy from someone using AI well and why knowing the desired outcome is the whole difference
• What your colleagues actually think when they spot unread AI output, backed by the numbers on trust and credibility
• The tells that give workslop away, from verbose answers to ignored brand guidelines
• Why “train your team” isn’t enough unless you define what the right way to use AI looks like at your company
• How making AI use visible turns a shameful shortcut into a useful handoff
• The one requirement that forces people to read the output before it leaves their desk

If your team is moving fast with AI and you’ve started to notice the quality slipping, this conversation names the problem and gives you several things to change today.

This episode is brought to you by Biggest Goal, the AI education company Alane Boyd and Micah Johnson co-founded.

Most leaders feel pressure and get stuck between ignoring AI and moving too fast. What works is hands-on training.

That's why Alane and Micah built Biggest Goal. Fifteen years of scaling their own companies taught them that adoption is a people problem long before it's a technology problem, so we train executives and their teams through workshops, cohorts, mastermind groups, and hands-on implementation support.

For more info about our AI programs, join our free community at your.biggestgoal.ai to get early access to episodes and learn about our upcoming events.

Thank you for supporting the podcast!

Transcript

Alane Boyd 0:00

Hey everybody, this is Elaine Boyd.

Micah Johnson 0:03

And Micah Johnson, and you're listening to Automate Your Agency by Biggest Goal.

Alane Boyd 0:08

In today's episode, we're going to talk about two terms that sound a lot like they should be in food manufacturing, but they're actually the newest terms in AI: meet proxy and work slop. Micah and I are going to talk about today what these mean and how to fix it in your org.

Micah Johnson 0:29

Elaine, these have got to be the two worst descriptions of things we've ever talked about on this podcast. Yeah, yeah. I know that, and that's even that's so much than them separate. Um, yeah, they're horrible terms, but I as much as I hate them, Elaine, I actually think they are so accurate.

Alane Boyd 1:01

There, there's a lot of truth to them for sure. And when and when you had this was actually something that was in your AI news brief that you send out in the community. And when I first saw that, I'm like, yes, this is exactly what is happening in organizations. Actually, Michael, before we jump into what it is, we've been experiencing these things because now teams, including ourselves, are so AI thinking forward. Hey, we want our teams using AI. And then what happens is, well, they use AI, but they are taking the thinking out of the process and just running a basically like running a skill or running a command or running it through AI chat, and they're delivering us just a result of what the AI said without looking at it. And that is them being a meat proxy.

Micah Johnson 1:51

Yes. So I'm gonna admit something, Elaine. I've been a meat pro like, I've done it. You you look at the output and you're like, good enough. Good enough. I'm gonna push this on.

Alane Boyd 2:04

I mean, I will say I've done it to my team embarrassingly enough. Like, hey, these are my ideas, and I just had AI put it all together for me, and then can you figure it out from here?

Micah Johnson 2:18

But you did something interesting there. So maybe you were a meat proxy and maybe you created work slop, but you identified that you used AI. Hey, I've worked with Claude, or hey, I worked with AI to assemble some of my thoughts, but I need help refining them from there. That's a really interesting point because I don't know that that would fully fall under this. And, you know, there's there's this gray area that we really have to figure out what is okay. And and I think every company might be a little different.

Alane Boyd 2:57

Yeah, it's true. And I'll take that, Mike. I'll say that I wasn't a meat proxy and delivering work slop. I was I had a few things in a few different places, and I thought AI would be the best way to get everything together for the team to go and execute. But but anyway, we when we are talking about meat slop, whether or not you and I are producing it are I called it again.

Micah Johnson 3:17

I think you called it meat slap again. That's okay. We can coin a new phrase here, Elaine. I think the the big thing that I that I want to point out before we go too far in is if you're listening to this and you're a leader, be aware that you could be the one doing this, not just your team or somebody on your team. I've seen it both ways, I've been guilty of it. And once you bring awareness to it, then that's one of the ways that you can really nip this in the bud.

Alane Boyd 3:47

Mm-hmm. Yeah, it's it that's true, Micah. Just like we're saying, you know, we've probably have delivered that in the past ourselves.

Micah Johnson 3:54

Mm-hmm. If a meat proxy is a person that takes AI and is just the proxy, so takes AI output and is the, you know, human meat bag to keep up a terrible, you know, naming convention here, uh, pushes it through. That's being a meat proxy, the output from that, which is just pushing stuff down the line in a process, is the output is the work slop. So just like you would hear like AI slop online all the time. This is the work product version of that.

Alane Boyd 4:30

Yeah, and you could see where this just continues to get pushed down through a department or cross-department. Well, the sales team produced this, sent it to marketing, they're doing, then they're turning it in and using. So it's like a continuation of just sloppy execution.

Micah Johnson 4:46

We've got some numbers that I think are really interesting that when it gets passed through and people don't identify that it's AI, and the person receiving it can immediately identify that it's work slot and go, this person didn't even read it. They just pushed it onto my plate, and now I have to deal with it. Um, these numbers are a cra are crazy because you have 53% are gonna um secretly feel annoyed but not say it out loud. 42% are gonna see the person that sent them that information as less trustworthy. Half of the people are gonna think those colleagues or leaders are less creative, less capable, and less reliable. And about a third of those people are not gonna want to work with that person again. And if you think about like how do we create a cohesive team? How do we keep things pushing forward? How do we create systems that scale? It's the automation, it's AI, it's all of this, but it's the soft skills too.

Alane Boyd 5:52

It comes down to, I mean, all those things are so true, Micah. Like I have felt felt all of those things as a recipient of work slump. Because what it means at the end of the day is now I have two choices. One, to send it back to you and have to manage you and tell you, hey, please update this because it's crap. Or two, I it now the work now falls on me. Neither one of those I want to do in my day, but now I'm set, I have to do this.

Micah Johnson 6:23

Oh, and if you're a task doer, if you're if you're assigned a task and you have to do that, you're a team member and you're getting this, you don't have the capability to go back up to your boss to manage them. Like you you're stuck, you're just stuck with it.

Alane Boyd 6:40

Yeah, people talking bad behind your back.

Micah Johnson 6:45

So, Elaine, what is your take on what is causing this?

Alane Boyd 6:51

I I think a few things kind of come to mind is one that we and I say we like it, I'll just use us for an example. We're always like, hey, let's use AI, let's make things easier for ourselves. And it's like when we were a kid and we like heard commercials and we believed like that commercial was truth. It's that we're taking those results that AI and believing, hey, they have more information than we might have. I'm just gonna take it. Or, man, that look sounds really nice. Or there's a million things on their plate, and they are just trying to get through as much as they can because they see their value in the number of things that got done that day. So I got an example of it because I feel like this could get confusing for somebody that's listening and go, I'm not an expert at this, and I use it all the time, and it's incredible. But that's the difference. You understand what the outcome is. And I'll I'll give an example of this. I love the legal plug-in. And it and I'm not a lawyer and I'm not trying to be one. I'm trying to get to an outcome because I've got 15 years of owning a company and experiences that I've learned that 15 years ago, if I would have asked Claude to put help me put together a contract, I would have got a shitty outcome. But I use I know what I'm looking for, I know what I'm trying to put together. I know when it spits something out and I'm like, actually, I don't like these terms. I want to change this. It's because I understand the outcome that I'm looking for. And I can put together the contract or the terms of agreement or whatever it is I'm working on because I do understand that. I can then send that to a legal expert to review, just like we would. And, you know, I'm not putting together work slab, I'm putting together something from my experience, from the things that I've come to understand, put that together and then send it to somebody to review. But I understood what that outcome that I was looking for was.

Micah Johnson 8:56

So this is such a great example, Elaine, because from a leader decision maker/slash strategic thinker perspective, this works. And this is why AI can be so powerful for somebody in your position making who is okay to make those decisions. Give that same responsibility to an intern. You give them the same contract, the same legal plug-in, and the same model that you're using, and you're asking for not as good of an output result, right? And they may get panicky and feel weird about this legal plug-in and go, I don't know anything about this. I'm just the damn intern. I'm just gonna go with what Claude is telling me to do. And, you know, everything with AI is on a percentage scale. So you might be getting it to like the 80 or 90 percent mark of what you're trying to produce for an output, and then you can get it to our legal team and they can take it from there. And it's shortened the back and forth. You know, you're we're not replacing lawyers, we're getting your ideas out to get in front of our lawyers so that they know exactly what you're trying to achieve with this contract or agreement or whatever it is, right? Or program. Um, but you give that to an intern and you don't give the intern enough context. Hey, help use Claude to review this. That might get it to the 40 or 50 percent mark. Now you're still getting it to the 40 or 50% mark with an intern. So, you know, one argument is you could delegate all that stuff that you're doing on the legal side of the lane to an intern to use Claude to use the legal plugin, which still gives it that extra skills, but you're still finally getting it to a legal expert. You're not going to Claude and saying, agreements done. I'm not running past, I'm not running this past the attorney. I'm just gonna leave it as is, right? And I think maybe there's so much to unpack here that we don't even have time for this in this whole episode. But the concept that I'm trying to say is there's different levels internally, but no matter what, it always ends in an expert.

Alane Boyd 11:09

Yes. And and that's where I think a lot of times a meet proxy, unless they just don't care about their job. I think the biggest thing when when somebody is a meet proxy is that, well, my boss asked me to do this. I don't know what they're looking for, I don't know what the outcome is, but I can go and have AI execute on this. So they're gonna do it. They don't look at it. Sometimes they don't look at it, but other times they don't look at it and they because they don't understand what it was that needed to be there.

Micah Johnson 11:39

Yeah. You know, my my favorite concept that relates to this, and this all gets back to managing people, not just AI, is the concept of you can ask somebody to do a task and they'll do the task, but they'll have no concept of why they're doing the task, why they're doing it at that moment, or anything around it. But if you ask somebody to do a task and you give them a reason to do the task, the purpose of that task, then all of a sudden that changes everything. And it's the same thing with AI and it's the same thing with people using AI.

Alane Boyd 12:18

It is. So how do you identify it, Micah?

Micah Johnson 12:22

I mean, one, right off the bat, if you're a domain expert and you get handed something by somebody else that just used AI and you look at it and you're like, WTF, this is this does not align with reality. You can instantly know somebody used quad. Somebody used quad.

Alane Boyd 12:44

I mean, everybody on LinkedIn knows if if you used AI or not. Some of it is obvious.

Micah Johnson 12:51

Some of it's super obvious. Yeah, some of it's just the writing, and then some of it will be the nuance of the materials that are output, the assets that are created. Mm-hmm.

Alane Boyd 13:03

I I think you know, when we're identifying it as when we're looking at the outcome, whatever was produced, and it's obvious that they didn't put thought into it.

Micah Johnson 13:14

Yeah.

Alane Boyd 13:15

They checked a box.

Micah Johnson 13:16

Yep.

Alane Boyd 13:17

And and I do find that people that love to do that are see themselves and find their value in the number of tasks that they do, that they tend to fall into that bucket of a meat proxy because they're just like seeing it as, oh, I'm doing my task, I'm doing my job, and not putting that creative or thought or strategic or effort into what does it look like to finish this?

Micah Johnson 13:42

Yeah. Yeah. I think another another um way to identify it is really long outputs, like super verbose long outputs for something that could take a couple sentences or a a summary or a couple paragraphs to figure out.

Alane Boyd 13:59

Man, I mean I love a few bullets.

Micah Johnson 14:01

Yeah, or or two bullets.

Alane Boyd 14:05

Tell me what you're gonna tell me in less than five words.

Micah Johnson 14:08

Yeah. And I think that covers like other things too. Like if you're trying to come up with a solution and somebody uses AI and they don't really understand why they're trying to come up with this solution. They're gonna, whether it's AI doing it or them working with AI, they're gonna try to cover every corner, every basis, every um aspect of it. And, you know, you look at the output and you're like, why are we talking about these eight other things? We only need to talk about these two things. That's what we're shooting for.

Alane Boyd 14:38

Yeah, I think that one's uh not something that I really think about often with the verboseness of AI, but you're absolutely right that most there's some people I can think of that love to talk a lot in writing and give me way too much to read, Micah. But but but for the most part, you know, AI is a lot more structured and talkative. And even if we say like cut out the M-dashes, like you know, those like tell all things that AI produced it, like take that out. And I'm just talking about like the language of things that that's an easy way to identify it. And then like the other thing is like if you have to say, Did you even read this? Did you look at what was done?

Micah Johnson 15:27

I you know, we see that a lot in like visual output and um like content or marketing stuff, Elaine.

Alane Boyd 15:34

I say marketing I a lot.

Micah Johnson 15:36

Yeah, that's another really clear tell where it's like, okay, it kind of hits the mark, but there's there's some real obvious stuff in here that like why would we say it this way? Or why didn't it use our brand guidelines? Or why didn't, you know, why aren't we promoting it according to the promotion that we have running right now, right? Like there's so many little tells in that.

Alane Boyd 16:05

Yeah, I had um she's no longer with us, but I had an assistant that would this wasn't marketing, but this was um using our call debrief skill. And we purposely have a human in the loop step when we run our call debrief skill, and it drafts an email. And I'm reading the email that she drafted for me, and I go, Did you even read this before you handed it to me? Because three of the things that it outlined we later on said we were not gonna do. So this shouldn't be included. So you are you reading the output or are you just doing the task?

Micah Johnson 16:43

Yes. Yes. And I think once you I think acknowledging and understanding that this is happening, whether it's you doing it yourself or people on your team are doing it, that's the first step. Then, like we talked about, like why, what causes it? I you you've got to identify that in your own company and your own team. Um I I think identifying it is probably the easy once you recognize, like, oh shit, this is obviously work slot.

Alane Boyd 17:18

I managing people and we are people are always like, is AI gonna replace humans? No, that the humans that put effort in they're not gonna get replaced. The effort has to be there. It doesn't replace us right now in this moment. And the people that are not producing work slop as a meat proxy are putting in the effort.

Micah Johnson 17:44

Yes. I I mean, I think about this every time we talk about this, Elaine, because I get I get passionate about it because I'm like, I know we've in prior episodes and as we were coming up through this whole AI era, it is like a little bit of a you know, oh crap moment. Like, this is some pretty powerful stuff. When you really start using it, when you start implementing it into your organization and and streamlining it, absolutely, this is an incredible tool to use. Now, tool is the correct term for this. Your train your team also needs to see this as a tool, not a partner, not a co-worker. Um, co-work is a tool, it's a tool for your human teams, the people and yourself to go, hey Claude, instead of me producing this whole thing, I'm gonna give you enough inputs, I'm gonna give you enough information, and I would like you to do what it takes somebody on myself or my team an hour or two. Please do that in five minutes. And then that hour or two that I'm saving, I'm gonna spend 10 to 15 or 20 minutes getting it right. That's not an error in the process, that's not an issue with the system. That is the freaking feature of the whole thing, is that yes, you have to do some work afterwards. Sorry, you gotta do some work. Like it, it is not replacing all the work, it's replacing the time-consuming, tedious stuff, and it's better at things than you at certain things, like cross-referencing large data sets and data analysis and summarizing and rewording and translation in fractions of the time. But those are the use cases that you want to use it for, and anything else is a misuse of the tool or not using the tool the right way.

Alane Boyd 19:42

So let's start talking about some ways that people and companies can fix it.

Micah Johnson 19:47

All right, but before we get into that, I'm not gonna get back on the soapbox, I promise. But I do want, you know, before we get into like some of the options on how to fix it, I do want to say if anybody is listening to this and going, oh crap, my team is a bunch of meat proxies, or I'm a meat proxy, and I want help solving this, this is what Elaine and I do for so many companies. So reach out to us at biggestgold.ai if you're interested in talking to us directly and want help with these types of problems that we bring up on this podcast.

Alane Boyd 20:26

Yeah, so you know, one of the ways that we talk about this all the time, whether we're talking about AI or not, but any tool, we always go back to train your team. And AI is no different. I see like so much, obviously, not the clients that we work with, you all are amazing, but we see so many other instances where teams are not trained. Nobody is talking about how to actually use it. They just get a subscription for their team members.

Micah Johnson 20:53

Yeah. And so I want to I I don't know how to frame this, Elena. Maybe you can help me, but training is a very broad word in this in this context. So giving somebody a license and saying, here's the functionality of co-work. Okay, that's better than no training. But then also giving people it's there's like a cult a company culture aspect to this with training. That's what's the right way to do it? What's the right way to use AI? Is work slop okay? Is being a meet proxy okay? And if it is at your company, then hopefully you stopped listening to us a long time ago in this episode because you wouldn't get anything out of it. But if it is, that's your company culture. And that is also what needs to go into training.

Alane Boyd 21:44

Yeah, that's a really good point. I think about um how right now, just how much in our company culture we're in the weeds on AI training. You know, what what development developing skills. How talking about usage and use cases. I mean, we have a Slack channel that we all of us put ideas in and use cases or ways that we used it or later latest updates. Like, and then we do a call with our team to dive in and we do things by department. Like we are very much in the weeds and learning from each other, learning from, you know, teaching other teams and figuring it out. Now, that doesn't mean that every team can operate like that, but having some type of assistance with your team, I think is required if you really want to get the efficiency gains out of AI.

Micah Johnson 22:42

Yeah.

Alane Boyd 22:43

Man, I mean, go back 15 years and the most relevant thing that I can think of is training your team on what is the single source of truth in your company. Because if you don't want work slop, guess what? It works from a single source of truth that everybody in your company also works from.

Micah Johnson 23:02

Yeah. The funny part is really how to solve all of this is best practices and how to build sys scalable systems, how to work with human teams. You put all that stuff together. Um, but training your team is huge. Another way is you've got to make this visible, like AI use visible. So it can't be shameful or it can't be shunned. Like your example earlier today, Elaine, or earlier in this episode was phenomenal because you said, hey guys, I'm using AI to help me shape my thoughts, create a prototype, do X, Y, or Z. Doesn't matter. But identify that you leveraged AI because that frames it so differently. Nobody, if if people are like, oh, you used AI for this, I don't know, then it must be, you know, if it's shameful or bad, then nobody's gonna admit it. And then nobody has that context and nobody has that information, and people react differently.

Alane Boyd 24:02

I also just think about like the the company culture as part of this. Like, there's one side that, you know, we're we're we openly talk about it with our team. Um, but also that in like the example that I gave, that I'm telling my team I used AI, I'm telling my team that it's crap right now, and don't take it for what I sent you. But that they had to feel like they could already say that about any work that I send them before AI.

Micah Johnson 24:33

It's very true. Very true. Um, and again, that's from a leadership perspective. So when when you're a leader and you want your team members to do this, they don't feel that they have that same capability to say, hey, next person in the line in this process or procedure, like, don't worry about this crap that I made, just make it better for me. That's definitely leadership speak. The answer is though, you have to define your process in a way that gives people the ability to know when it's okay to use their judgment and assign checkpoints in your process that is exactly defining hey, this is when we're going to review, validate, use human judgment, et cetera, in this process. And it's not okay to just push it forward because you don't know. At some point, if you don't know, you have to ask.

Alane Boyd 25:28

You know, another one that comes to mind is thinking about like the number of things, the the expectations around the number of things that need to get done in a day, especially now that you have AI as a tool. Because what I've noticed is if we're talking about like a uh a co like using cowork, where it's it that's an individual relationship for the most part. You're using a skill, it's to help you day to day, then especially in the beginning, it takes a little bit to fine-tune exactly what you're trying to tell co-work or Claude to do, maybe developing the skill. So there's even some extra time that can take up front, but to save you hours of time later on. So in the beginning, it may not feel like it's really saving any time to where you can get more done in a day.

Micah Johnson 26:20

Yeah. And I I would even say on the opposite side of that, culturally, there has to be expectations set. If you're rewarding the person who's putting out volume and not quality, that's a company culture issue. The people who are putting out quality need time. They need time to think, they need time to leverage strategy, they need time for focus. People that are putting out volume, you know, and it and this really depends on the roles and the responsibilities for these individual people. But generally speaking, putting out volume means you're shortcutting something. And AI is a beautifully easy shortcut to take. So in most cases, like you're saying, Elaine, you don't want to reward volume, you want to reward quality.

Alane Boyd 27:04

The amount of things that I'm touching in the day may be the same amount, but the finish line I'm getting to is faster.

Micah Johnson 27:13

Oh, yeah. I like that. I like that. So over a week, you're getting more things done, but you're still working on the same number of things per day.

Alane Boyd 27:24

Love that's really trying to think of how that how I'm trying to think of that. But it's like so many things that it would take me longer to get to the finish line that I really couldn't get done, or that I'd have to keep working, you know, a little bit more each day on, or something like that. It's now I can get it done in in a day rather than four days, or maybe it gets on my backlog and I really need to get it done, but I just don't have the skill set to get it over the edge, or you know, whatever it might be.

Micah Johnson 27:51

I mean, design back and forth with validation, all of those things are what locks up a lot of managers and leaders and directors days because it has to go to somebody that can produce it. And they with AI, you can go back and forth much, much more quickly to get to the output. I really like that concept. Um, one of the things that we've started experimenting with, Elaine, is having people give a human written summary on the output, which means they have to read the output. And if the output is too verbose, they have to work with AI to consolidate it down and ultimately get to a point where, hey, I've read this, I've I understand the output. I'm not just blindly pushing it forward to be somebody else's responsibility. And here's my take on it. And I love this one because you can have AI give its take, fine. You can have it give an improved take by giving it better information and not just running off of training data. Absolutely. But then if you do what we've talked about today and make AI use visible and write the summary or your take on it, now you've got like this impressive like, this is what AI says, and here's my expert opinion.

Alane Boyd 29:10

One's one thing that I thought that Yanelle does on our team, she's such a rock star. She'll she uses AI to help put together my speaking proposals and things. But then she goes through and comments on the document on things that she's like, hey, this is a concept I'm working through, but I want your you to look at it and give me some of your words around it. So she's actively going through it and calling out the areas. So I know by the time I'm looking at that product, she's already gone through it and made edits herself before she's even commenting because she's actively reading through it.

Micah Johnson 29:46

All right. So the bottom line, I would say, Elaine, is that we've got this amazing tool called AI. It has when used correctly, you can produce so much, so so much more that's still quality and create all these assets extremely quickly. But it doesn't stop you from needing somebody on the team, whether that's yourself or somebody else, to use this tool with the understanding of why they're using the tool and what they're using it for, and that the outcome of the tool is hitting that output that's desired in the first place.

Alane Boyd 30:28

Micah, have you ever read the book R uh Radical Candor?

Micah Johnson 30:33

I don't think I've read that one.

Alane Boyd 30:35

It's it's been several years since I read it, but I I every time we talk about meet proxy and work slop, I think about this book. Because at the end of the day, you have to have radical candor. You have to say this is work slop or describing it if you don't want to say the word work slop, but saying this is not okay to just send me this, you know, stuff. Like you have to have these conversations with your team, which is still part of managing people.

Micah Johnson 31:11

Yes. Yes.

Alane Boyd 31:15

So if you haven't read that book, read the book. Mike, I'm gonna put it on your to read list.

Micah Johnson 31:20

Perfect. I will read that. All right, everybody. Well, thanks for listening to today's episode. Next episode, Elaine, we are going to find less disgusting words to discuss. So stay tuned. Um, but every episode on this on this channel or in this podcast, we talk about a bunch of stuff. AI. It really boils down to how do you build a scalable company? How do you optimize your operations? And if you want help with any of this, go to your.biggestgoal.ai. That's our free community. We have masterclasses, we have co work starter kits. We're running tons of programs right now. So reach out to us with any questions. Reach out to us with anything that you are struggling with on AI. We'd love to hear from you uh and keep listening. Thanks, everyone.