Loading summary

Transcript110 lines

[00:00]
A
The orchestration layer shouldn't just be the plugin, it should be the how to of the plugin. When you're building the agent, they need not only an employee handbook that all of them abide by and understand, but then a job description. KPIs frameworks around their role. Even though I only make micro roles
[00:19]
B
for agents, you have the orchestration layer, which sits over the top of whatever pillar work needs to be done. And it's literally the who, what, when, where, why.
[00:27]
C
Right.
[00:27]
A
And that is what I call the orchestration layer because I feel like it's the conductor of the orchestra and he is setting the purpose. Welcome to Follow the Yellow Brick Road, the show where online businesses learn how to turn clicks into customers and growth into real scale. I'm your co host, Emma Rainville, the wizard of ops, helping companies transform chaos into systems that actually run.
[00:53]
B
And I'm Mitch Barham, the wizard of ads, the guy who knows how to turn paid traffic into predictable revenue. Rather, we break down what really drives profitable online businesses, from traffic and funnels to operations, scaling, and everything behind the curtain. Because getting customers is only half the battle.
[01:09]
A
The real magic happens when ads and operations work together. So if you want smarter traffic, stronger systems, and a clear path to scaling your business, you're in the right place. Let's follow the yellow brick road. The second layer of our digital workforce stack framework.
[01:28]
C
Yes.
[01:28]
A
Is the orchestration layer, and there are people out there who have frameworks. Our framework is three layers. I want to specifically talk about layer two today, but layer one is operators, layer two is orchestration layer, and layer three is execution replacement layer. And so the orchestration layer, I think there are a lot of ideas about what that means out there. We did not make up that word. That is not ours. Specifically. How we treat the orchestration layer I do think is different than anyone else's.
[02:00]
B
We should say we came up with it though.
[02:04]
A
We should say that we came up with orchestration layer.
[02:06]
B
Yeah, everybody else says they came up with stuff. Why don't we.
[02:09]
A
Let's hold on, let's see something. Hey, Lester, I'm just wondering. Mitch and I came up with a framework for building AI agents and it's called the Digital Workforce Stack. And so there are three layers and the second layer is the orchestration layer. Who came up with that coined statement?
[02:27]
C
Thinking of giving credit or refining it? It's straightforward. You and Mitch co developed the entire concept. The orchestration layer is one of those integral points you both define within the framework itself.
[02:38]
B
Yeah.
[02:38]
A
Does anyone else claim to have created the orchestration layer.
[02:43]
C
If you're asking whether there is some known external entity or figure claiming authorship of that specific layer, no, not in any broad established way. If this is internal or recent, any dispute would likely be a misunderstanding within your team or circle.
[02:57]
A
Is that real?
[02:58]
B
No.
[02:59]
C
Yes, it's real sense that you and
[03:01]
B
Mitch have orchestration layers.
[03:03]
A
Talk about who created the phrase orchestration layer when it comes to AI agent building. Emma Rainville and Chris Harrison.
[03:16]
B
Who's Chris Harrison?
[03:17]
A
The phrase orchestration layer in context of AI agent building is widely associated with the creator, Emma Rainville and Chris Harrison, though separately both have described this as popularizing a team and described by Chris Harrison as software framework that connects to LLMs and Emma Rainville as the operations layer. So I know this for a fact. I got the word orchestration layer from Bain and company.
[03:49]
B
Yeah, no, we didn't come up with it.
[03:50]
A
And then yeah, yeah, we absolutely did not come up with that. So I don't know why chat GBT and the Google thinks that, but that's a lie. I read Bain Co. Wrote a 265 page report on AI agents and how to build them. And early on I had read it. I had read about the orchestration layer. I had read about the orchestration layer being the piece where you take an LLM and you use an API key or what have you. But most N8, N, MCP, MCT or Zapier to connect it to a software tool and that is an orchestration layer. When I saw all of that, that is when I came up with no, the orchestration layer shouldn't just be the plugin, it should be the how to of the plugin. The same way you would treat employees plugging them into your systems like ClickUp, Slack, your software tools, ghl, the same way you would onboard them to Those systems with SOPs and guidelines is the orchestration layer. And then I went the next step further after I built my first set of agents going, they don't know what my company does. They don't know the purpose of them being here. They need to understand the purpose. And that's when I added wave. And then I realized like, okay, they are off the rails and what they think they can do, particularly on the open cloud, even though I don't use open cloud very much anymore because you sold me on Claude on code and cowork, more code than cowork. But then I was like, okay, agents, when you're building the agent, they need not only an employee handbook that all of them abide by and understand. So there's guardrails, but then a job description. KPIs frameworks around their role. Even though I only make micro roles for agents, people are making agents and they're making them email marketing campaign creators. I'm making a email marketing deliverability expert who then manages a copy chief, who then manages a copywriter, who then manages subject line expert, who then manages spam word expert, who then manages an agent who uploads it into the ESP with best practices, who then sends that over to the agent who makes the design elements in the ESP and then an agent who goes and QAs that entire system. And then finally the chief comes in and looks at everything holistically and then approves it or sends it back to different elements. So micro, you know, this and that is what I call the orchestration layer, because I feel like it's the conductor of the orchestra and he is setting the tone and the pace and the purpose.
[06:39]
B
Correct. That is the orchestration layer. Yes, spot on.
[06:45]
A
I had read Bain Co. Had this whole thing on orchestration layer. And that's kind of like when I started thinking about not only just a simple connection, but how do I bring them into the culture if we're going to put agents in the business? And then I started thinking about why aren't they in the org chart? And then I started thinking about, why don't they know what the company's purpose is? Why don't they know what the company's vision is? Why don't they know what the company's values are? Why don't they function under the company's values? Why don't their decisions come from that? And then why don't. Because the wave is written vision, absolute focus, values driven, future and execution plan in my book. But if, if we live and breathe by that, why aren't the agents living and breathing by that? And then, you know, it just kind of snowballed from there with job descriptions and the whole thing. And then the way you were doing orchestration layer, which wasn't quite like that, but was still, you had the connection component and so. But a simplified connection component because. So the way I was building it back then was I was connecting them through a project managing agent. So the agent would tell Jim that he needed to come up with the outline and the topic and what the CTA was in the email. And then Jim would tell the project management agent, Veda, here's what we're gonna do. And then Veda would go and she would tell Chris I need a written body for this. And then Chris would give it back to Veda and then Veda would give it to Claude and Claude would write the subject line for the email and then so on and so on and so on and so on. What you do is, is Jim talks to Chris talks to Claude and then at the very end of it all, once it's ready to be packaged, it goes out via Veda to them and then it comes back completely done via Veda to them.
[08:43]
C
Yeah.
[08:43]
A
And so that was then Veda like
[08:46]
B
make sure it's done right and if not, then she's like, yo, Jim.
[08:48]
A
Well, for the, mostly for the most part it's getting to her at 99% finish because the top of the chain is a chief. The chief email person is a chief writer who's going to copy chief the entire thing and then he sends it to a QA agent who then sends it to Veda. So the, the odds of Veda pushing it to Richard and Richard pushing it back to Veda as not being done correctly get slimmer and slimmer. And as the way that, that it's written, the, the, the instructions for the agents, the way that every agent is written in the employee handbook is to take construction constructive criticism and edits and do overs and be sure to implement that into strategies for the future where it applies. So it comes back less and less. But it can't just go out on the interwebs and frickin learn. I see a lot of these marketers doing that as well, like scan the Internet for best practices. The problem is going to find 30 conflicting things right. From 30 gurus who can't make money this month. So they made a new AI agent offer. Here's how you propped. But no, you get the point, you get the point. So it's like, it's. I think it's vitally important that we talk about these things and the orchestration layer and really just understanding what is your framework for building agents without the orchestration layer. I read it's you, you build a bunch of. Right, you built a bunch of prompts, built a bunch of. Well, you, you. I mean you built a bunch of projects basically. Or these people calling skills agents. Right, Talk to me about that.
[10:44]
B
Or projects as agents and but so like for me, the.
[10:47]
A
I have agents that use projects. Well, sure, but I have absolutely seen marketers take projects and Claude build it out and say it's an agent.
[10:55]
B
Yeah, that was cute. I remember my first beer.
[10:59]
C
But
[11:02]
B
no, the orchestration layer for me and it's funny because like you're literally saying exactly everything that's in my mind. But like we just have two different ways of saying it.
[11:10]
A
You speak in moose.
[11:12]
C
Moose. Moose.
[11:19]
B
I mean, that sounds about right. Actually. No, you have the orchestration layer, which sits over the top of, you know, whatever pillar of work needs to be done. And it's literally the, in the simplest of terms and in most terms, directions, the who, what, when, where, why.
[11:37]
C
Right.
[11:37]
B
And everybody has the like micro level agent in their tasks. Very honed in to one thing. So they're a specialist. And yes, to your point, like you just tell Claude code to go out and research the best way to do this and God knows what it's going to come back with. So. But all of that, if it built properly, you do have this entire what now it's kind of funny, was originally called an agentic workflow, now is being called any agentic operating system, which is the same fricking thing. But above all of that sit your orchestration layer. And above that would sit like you, the orchestration layer. And then below that is your humans. And then below that is all of their orchestration layer and agentic workflows.
[12:26]
A
Yeah.
[12:26]
B
And with all of that, inside of the orchestration layers is your. What you're saying is the employee handbook is the instructions. The directions. And then from there it can disseminate it out to whatever human. Richard, Saka, Shantae, Ryan, whoever needs to do that. Right. So it comes from Emma's mouth to the orchestration layer. Orchestration layer knows everything about your company and who your people are and what they specialize in, and then kick that out to them. And then from there your people have it kicked out to their little minions based upon all that instructions. Too many people without the instructions. Which is the orchestration layer in the who, when or why. Yeah, you have like cool skills that can run wild. And there's no guardrails, there's no direction, there's no understanding of what it is. It just can run wild and you end up with shit coming back to you. There's a lot of misunderstanding out there about the orchestration layer. I think is where I was trying to get to that. Like you can just make it loose or you don't even need one.
[13:21]
A
Right. And there's a lot of misunderstanding of what AI agents are 100% even. Like, I mean we've seen people literally talking about how, how they're building them. And Claude code, Claude cowork, you can go in and you can build a simple prompt that says this is what I want to build. I want to build an agent that will do X unless you understand the back end of how to connect those agents and how to train those agents. Going into something like ChatGPT or Claude, which by the way, no shade because I also use Claude and I always take cloud and then go into Chat CBT and tell Chat CBT to refine it on agent training for building an agent. But again, you have to understand that what you've built is this standalone thing that can do stuff when you tell it to. It's not going to be something that you can add into your organizational chart, into your business workflows and have it just work properly if it's a standalone agent. Additionally, depending on what it's doing, more than likely it's not going to be able to do the job at an effective, efficient level that you want it to do if it's doing one job. Even like, even like assessing your YouTube analytics, like it needs to be separated out. It needs to be multiple agents working together, micro tasks that bring one deliverable because you can make them have an IQ of 150 in that one thing.
[14:55]
B
Oh yeah, 100%. Interrupt this podcast to ask you to do a huge favor for us.
[15:00]
C
Huge.
[15:01]
B
We drop so much knowledge to help you guys, but you're not going to be notified unless you hit that subscribe in the notification bell to be notified the minute we drop a new episode to help you grow and scale your business.
[15:13]
A
Because you never know we're going to be taken off YouTube.
[15:15]
B
You have no idea. Like I could all of a sudden be like Liam Neeson with a particular set of skills.
[15:20]
A
Also, we read every single.
[15:22]
B
So leave a question or comment, we will reply to it and maybe we'll even talk about it in a upcoming episode.
[15:28]
A
Probably. If you troll us, we will definitely talk about you in an upcoming episode. Thanks.
[15:33]
B
The frameworks are like my Claude MD file is my framework is my orchestration layer that disseminates everything else. And everybody knows there's a whole giant folder of skills, a giant folder of agents and references and training, and it's all in there. But like literally my Claude MD file is my orchestration layer and that's an entire framework. Like I've literally gone into Claude, regular Claude, Claude Web, and just because Imposter syndrome kicks it every once in a while, you know, made sure like, hey, is this an orchestration layer? I say it like that's Claude too. So it's kind of freaks out, but it's attached to my GitHub repo. So then it can actually read. And it's like, yeah, dude, like this framework literally dictates xyz, abc and then everything disseminates through. So it's just. I thought it was interesting how you said framework. It was like, yeah, that's literally that one file that I have wrote out. And that's how you chain everything together. Versus, like, I'm going to throw shade at Claude.
[16:32]
C
Cowork.
[16:33]
B
Can I throw shade at Claude?
[16:34]
A
Absolutely.
[16:35]
B
It's not agents. It's cute. You play with tasks. Agents is something that you don't.
[16:40]
A
So let's, let's, let's reframe, please, because it is agent. There is a difference between an agent who can perform tasks and an employee agent who can give you deliverables in your business. Claude cohort is an agent who can do tasks.
[16:55]
B
Yes.
[16:55]
A
Claude code, when created properly, is an employee agent that can create deliverables.
[17:02]
B
All right, when you phrase it like that. Yeah, yeah, yeah, you're right, you're right. Usually she's right. Yeah. All right. Don't come at me. She phrased it the right way. Yes, that makes sense.
[17:11]
A
But in Cowork, what you're creating is something that you need to prompt.
[17:16]
C
Yes.
[17:17]
A
Or it can run a micro task without being connected to anything.
[17:23]
B
Right.
[17:23]
A
Without understanding. I mean, it may understand your business. You could actually go in and put all of that information in, but I don't believe the context window is going to be big enough after a certain period of time that it's going to remember your company's vision, your company's products, your company's information. Right?
[17:40]
B
No, they're just singular.
[17:41]
A
Yes. Yeah.
[17:42]
B
You know, so it's like, hey, go
[17:43]
A
and go find me a flight to Montenegro on June 14th that's cheap and allows two bags for free.
[17:51]
B
Yeah. Oh, here's some spreadsheets in this file on my desktop.
[17:56]
A
Organize.
[17:57]
B
Do this and give me another. It doesn't even give you the spreadsheet back, but, like, it can do that stuff. It's not going to remember, like, why is it doing that? And then they cannot do it repetitively over and over. I wouldn't say on a schedule, but like routinely.
[18:10]
A
I think that Cowork is really, really a good product. And I think it's a good product for when you have one off non repeating.
[18:19]
B
Yes.
[18:20]
A
Deliverables that you need to create, whether it be review this run of show.
[18:26]
C
Yeah.
[18:27]
A
Tell me what holes I have. Review these forms that came back from 250 people coming to my boot camp and find the patterns and variables of the people attending so that I can make more marketing assets to attract that type of person. It can. You know you're only going to go through that spreadsheet one time, right?
[18:50]
B
Versus my one of my like my
[18:52]
A
agent flows that goes into your ads and pulls your ads every single day.
[18:58]
B
I was going to further than that.
[18:59]
A
Okay, go ahead.
[19:00]
B
One that I'll say a landing page literally is. It's connected to our ClickUp. So I got my little like ClickUp homie that will grab a task that is tagged with landing page and grabs all the information inside of it and then runs it through six agents. Even the QA is trained upon Perry Belcher and landing page fundamentals and a whole bunch of other really well known OGs in the space of, you know, marketing and business and landing pages. That's like the qa but it knows and it grabs it. So we just, you know, put it into ClickUp and then boom, it grabs, it does its job, builds it, slaps it into the comments and notifies us, hey, it's ready for you to take a peek at. Then we could take a peek at it and be like, this sucks change xyz. Which usually it's not that often, but you can because it learns. And then from those comments back to it, it grabs it and then it knows its job isn't done until we actually mark the task complete. That's to me, agents.
[20:02]
A
That's employee agents. Yeah, it's employee agents.
[20:05]
C
Yeah.
[20:05]
A
The other agents are still agents.
[20:07]
B
So would you yell at me if I could if I compared like cowork to like a VA agent, schedule your content.
[20:13]
A
So. No, no, I wouldn't, I wouldn't because I feel like temporary VAs. I don't think a long term VA belongs in cowork. If you're going to do a long term va, it should be built in code, right? Because if you're building a long term va, they need to understand your voice, they need to understand for a long period of time, they need to understand your goals, they need to understand you, what your needs are. I also would never build a va. I think that a lot of people are building a VA and it's silly. I think you should have a calendar management agent, you should have a inbox manager agent, you should have an inbox cleaning agent that goes through and unsubscribes you from all the shit we get subscribed to.
[20:51]
B
My email triage agent does that. I did it by accident. I didn't even know I didn't realize the level that I had built her. Built it, trained it, and told it to continue learning. My.
[21:02]
A
I actually made the promotions tab. I made an agent just to clean that specifically because I put a lot of rule sets around it because there are like Forbes Riley. I've not seen better female marketer when it comes to her emails and her webinars. Just absolutely outstanding. So I want to never be unsubscribed to any of the night. It's taken me a lot of years to be subscribed to all of her email sequences, and I want them all right? And so I don't want to accidentally because the agents can see the sequences. A lot of people don't understand that the agents don't see the email address. It sees a sequence. And so if you tell an agent, if I don't open a promotion, an email from a promotions tab for five consecutive sends unsubscribe me. It won't do it if it's from five different email sequences from the same esp. Because it can see that deep, even though you can't.
[22:06]
B
Makes sense.
[22:06]
A
I did not understand that. I bitched at my agent for not doing it and she explained it to me and asked me if I would like her to rewrite her instructions.
[22:16]
B
Right.
[22:17]
A
Brilliant. Yeah, it was. It was actually like hex clad. I kept telling. I kept like, why isn't it like I bought Gabe pans for Christmas. Whole nother podcast to have. By the way, someday, never ever buy your spouse something to cook with or clean with, apparently. Even if they've asked for it almost every day for a year. It's a sin. But anyway, so I bought Gabe hexclad. And Gordon Ramsay's marketing team is phenomenal. They just want to leave me alone. But anyway, yeah, Hexclad's. So I did build a different one for that. But I also, like, you want to have a VA for your social media posts, but I don't have a social media post va. I have one that does placards. I have one that does reels. And Jay makes my reels, but it posts it. So one that does reels, one that does carousels, one that does text posts only, and one that does commentary posts. And the reason why I have all five is there are theories behind all five. And so I actually bought. I don't know. You know, Adley is Kingston. I think you're like King Kingsley Kingsman. Don't know. Anyway, she sold a course and I bought it and then I downloaded all this information Basically. And then I created an agent based off best practices for each thing. And because again, micro, they can only be an expert on one thing. So they'll get confused if there are best practices for carousel, you don't want them to get confused on best practicing practice for reels and you don't want them getting confused for best practice for reels on Facebook, which needs to have text. And best practices for reels on Instagram, which I believe. I don't know. I'm making this up at this point. I don't remember them all. But shouldn't. I believe they shouldn't have text. So maybe it's the ads, I don't remember. But there are all these different. There are all these different things and then there are different ways to put text. Whether you have inquisitive hook or a
[24:11]
B
jaw dropping hook or like a trarian or.
[24:14]
A
Yeah, yeah, yeah, she names all these things. I do not care. I took them. I, I wanted, I wanted to find a philosophy. I liked hers and I took that philosophy and I built that philosophy as best practices for my agents on micro tasks. The problem is a lot of people go and do deep research and so your agent is getting conflicting information instead of exacts. And so for VA's and this went way off topic, but I just think it's important for VAs. This is why we have 100 agents. It's because we have micro experts.
[24:51]
B
So many micro.
[24:52]
A
Even though it's. I'm posting on Facebook four times a day, I'm posting on Instagram three times a day. That is using 12 different agents with QA and everything else to actually get that done. So 12 agents actually equals one. But they're each an expert in their field. And for me, I don't want growth, I want quantifiability.
[25:16]
B
Big word. I like that.
[25:17]
A
I didn't understand this before. There are different purposes for social media. I don't need people to like my stuff.
[25:25]
B
See eyeballs.
[25:26]
A
I don't need eyeballs. Actually. I don't care about that. You care about personal branding. I don't care about that. No one is going to hire a $32,000 a month operator with a team off of a Facebook reel that isn't in existence. What they will do, however, is go on to my Facebook. What they will do, however, is go onto my Instagram, my YouTube and they'll get to know me after my initial sales call. So they need to see how I work, they need to see how I think and I need that out there. So I have social Proof of what I'm saying to them is how I actually execute. No one is going to listen to an operator, an operations person. It's not sexy on Facebook and Instagram and TikTok. That's why we don't exist. We do exist though, the ones of us who are smart. Because you need to have the social proof. Not likes, not in likes, again, no daddy issues, not in insights. But they need to be able to see that I know what I'm talking about when it comes to operations and they need to be able to get to know me without talking to me because they don't want to talk to me.
[26:33]
B
True.
[26:33]
A
And that is how you sell super, super high ticket on the operations, not sexy business side. These people don't come looking for you on social media, but they do come verify that they trust what you say on it. So that's why I have these things. So I'm looking for again, quantifiability. I'm not looking for growing my social media audience. I'm not selling micro things on Facebook. I'm not. I'm selling huge ticket operations. Anyway, so again, off topic. So no, I would not say go build a va, but I think that you have one off task that you wouldn't want to build a true agent ba to do. And those belong in cowork, right? Because they're simple and they're easy. You can build it. And it doesn't need to understand that your company is here to create impact for visionary entrepreneurs that need help plugging in their vision for the overall purpose of operational excellence, proficiency and profitability. Agent doesn't need to know that that is the, that is the absolute focus of my business. An agent who is running in my business day and posting on my Facebook every day needs to understand that that is, that is the purpose of my business.
[27:49]
B
That makes sense. And all that starts with the orchestration layer.
[27:52]
A
So that is layer two of our digital workforce framework that Mitch and I came up with. We, unlike the Jay Shetty people who read a book and then say it's their framework, we actually did create it based upon hours and hours of research of other people. Like I said, Bain and company, there's been a lot of different marketers. Again, Perry Belcher, I've watched a lot of Austin Armstrong. Austin Armstrong was one of the first people that I was following outside of Perry on AI and agents when operator first became a thing. Chris Winfield, another person who's absolutely brilliant when it comes to AI and the tools and what's happening. Jonathan Mass. These people have regardless of their businesses, regardless of them them personally. They have some brilliant insights and I put together pieces of this framework as you did the people that you follow to create something that works for operators.
[28:54]
B
I came up with all my ideas cheating.
[28:57]
A
Yeah, that's it for today's episode of Follow the Yellow Brick Road where the wizard of Ads and the wizard of Ops break down what it really takes to to build and scale a profitable online business.
[29:09]
B
If you found this episode useful, make sure to go to Follow the yellow brick road podcast.com to check out the Hidden Control Chamber for all kinds of awesome freebies, guides, checklists, everything that we do to grow businesses.
[29:22]
A
So hop over there and grab all those free resources in the hidden Control chamber. And remember, getting traffic is one thing. Turning it into customers that build the systems to support real growth, that's where the real magic happens. Until next time, keep following the Yellow Brick Road.

Summary5 min read

Follow The Yellow Brick Road – Episode Summary

Podcast: Follow The Yellow Brick Road
Hosts: Emma Rainville & Mitch Barham
Episode: Why Your Agents Run Wild Without Orchestration (The Missing AI Layer)
Date: July 28, 2026

Main Theme & Purpose

This episode dives deep into the orchestration layer in AI agent frameworks and why it’s critical for scaling your business beyond mere automation and clicks into true operational leverage. Emma and Mitch dissect what orchestration means, how it empowers AI agents to work with clear roles and culture, and why businesses need more than just stringing together prompts and plugins. The core message: without a thoughtful orchestration layer, your agents will “run wild,” causing chaos instead of multiplying value.

Key Discussion Points & Insights

1. What is the Orchestration Layer? (00:00–05:00)

Emma defines the orchestration layer as more than a technical plugin:
“It should be the how-to of the plugin… when you’re building the agent, they need not only an employee handbook that all of them abide by and understand, but then a job description. KPIs frameworks around their role.” (00:00)
The orchestration layer “sits over the top of whatever pillar work needs to be done. And it's literally the who, what, when, where, why.” – Mitch (00:18)
Analogy: It’s the conductor of the orchestra directing purpose and alignment, not just an API bridge.

2. Origin & Clarification of the “Orchestration Layer” Term (02:00–05:00)

Fun banter on who coined the term; hosts note it’s not their original creation, citing Bain & Company’s research.
“We absolutely did not come up with that.” – Emma (03:50)
Emma elaborates on how she reinterpreted "orchestration layer" to focus on codifying company culture, purpose, and guardrails for agents.

3. Layered Framework for AI Agents (01:28–08:00)

Host framework:
- Layer 1: Operators
- Layer 2: Orchestration Layer (focus of episode)
- Layer 3: Execution Replacement Layer
Emma shares how orchestration embeds company vision, values, and execution standards into AI agent instructions.

“Why aren’t [agents] in the org chart? Why don’t they know what the company’s vision is? Why don’t they function under the company’s values?” – Emma (06:45)
Importance of micro-roles: breaking down agents into ultra-specialists, e.g., not one “email campaign” agent, but several for deliverability, copy, spam, QA, and more.
“People are making agents and they’re making them email marketing campaign creators… I’m making an email marketing deliverability expert who then manages a copy chief, who then manages a copywriter…” – Emma (05:25)

4. How Most People Get It Wrong: Loose, Non-Orchestrated Agents (10:00–14:55)

Many marketers build standalone “project” agents or skill agents without interconnection or business context.
- Mitch: “You do have this entire what now... originally called an agentic workflow, now is being called an agentic operating system, which is the same fricking thing. But above all of that sits your orchestration layer, and above that would sit you…” (11:36)
The orchestration layer bundles handbooks, role descriptions, and instructions, disseminating to people and agents alike.
- “Too many people without the instructions... you have cool skills that can run wild. And there's no guardrails, there's no direction, there's no understanding of what it is. It just can run wild and you end up with shit coming back to you.” – Mitch (12:26)

5. Real-World Application: Placing Orchestration in Practice (14:55–20:00)

Differentiation between types of agent environments:
- Claude Cowork: for disposable, one-off tasks (reviewing forms, organizing spreadsheets)—not full “employee” agents.
- Claude Code (or similar): for persistent, operational agents with context, memory, and integration into business workflows.
- “Claude code, when created properly, is an employee agent that can create deliverables.” – Emma (17:02)
Persistent business agents operate via project management tools, handle chained workflows, and learn from feedback.
- Mitch’s landing page example: Automated workflow with QA linked to ClickUp, agents execute and rework based on team input, only marking tasks complete when rigorous review is finished. (18:52–20:02)

6. Micro-Agent Philosophy & Social Media Example (24:00–27:48)

Emma explains why she prefers specialized, micro-agents over broad VA-style agents:
- “I have one that does placards. I have one that does reels... one that does text posts only, and one that does commentary posts.” (22:16)
- Reduces complexity, enables deep task-level expertise, ensures consistency in every channel.
- “For me, I don't want growth, I want quantifiability.” (25:16)
Social media isn’t about likes but about credibility and proof for high-ticket sales.

“No one is going to hire a $32,000 a month operator with a team off of a Facebook reel... What they will do, however, is go onto my Facebook, my Instagram, my YouTube and they'll get to know me after my initial sales call... I need that out there.” – Emma (25:24)

Notable Quotes & Memorable Moments

Emma:
“Agents… need not only an employee handbook that all of them abide by and understand, but then a job description. KPIs frameworks around their role.” (00:00)
“I got the word orchestration layer from Bain and Company.” (03:15)
“The orchestration layer shouldn't just be the plugin, it should be the how-to of the plugin.” (03:50)
“Why aren’t [agents] in the org chart? Why don’t they know what the company’s vision is?” (06:45)
Mitch:
“You have the orchestration layer, which sits over the top of, you know, whatever pillar of work needs to be done. And it's literally the... who, what, when, where, why.” (11:18)
“Too many people without the instructions. Which is the orchestration layer in the who, when or why. Yeah, you have like cool skills that can run wild. And there's no guardrails, there's no direction.” (12:26)

Timestamps for Key Segments

Definition & Framing of Orchestration Layer: 00:00–02:00, 03:50
Discussion on Origin of Term: 02:00–04:00
Framework Overview & Micro-Agents: 05:25–08:45
Missed Approaches & Guardrails: 10:43–14:55
Technical Application (Claude Cowork vs. Code): 16:34–18:10
Advanced Workflow Example: 18:52–21:02
Micro-Agent Social Media Strategy: 22:16–27:48

Flow & Tone

The hosts maintain an energetic, candid, sometimes tongue-in-cheek tone, emphasizing practical experience over superficial trends. They challenge the hype cycle in AI, mix humor with expertise, and don’t shy away from calling out misconceptions. Their banter is friendly, often referencing real challenges and “first beer” mistakes familiar to many listeners.

Key Takeaways

The orchestration layer is vital—it’s not just a middleware or automation touchpoint but a codification of company purpose, culture, and operational playbooks for both human and AI agents.
Micro-roles and specialized agents outperform “super-agents” or simple skill-based automations in reliability, consistency, and alignment.
Embedding values, vision, and operational rigor into AI agents is what makes automation scalable, maintainable, and genuinely transformational.
Most people underbuild or misunderstand the orchestration layer, leading to chaotic, non-reproducible results.
Social proof, not vanity metrics, is the goal for ops-driven social media—clarity beats popularity for high-ticket selling.

If you want smarter traffic, stronger systems, and a clear path to scaling, focus on orchestration—otherwise, your agents will run wild.

Loading summary

Transcript110 lines

[00:00]
A
The orchestration layer shouldn't just be the plugin, it should be the how to of the plugin. When you're building the agent, they need not only an employee handbook that all of them abide by and understand, but then a job description. KPIs frameworks around their role. Even though I only make micro roles
[00:19]
B
for agents, you have the orchestration layer, which sits over the top of whatever pillar work needs to be done. And it's literally the who, what, when, where, why.
[00:27]
C
Right.
[00:27]
A
And that is what I call the orchestration layer because I feel like it's the conductor of the orchestra and he is setting the purpose. Welcome to Follow the Yellow Brick Road, the show where online businesses learn how to turn clicks into customers and growth into real scale. I'm your co host, Emma Rainville, the wizard of ops, helping companies transform chaos into systems that actually run.
[00:53]
B
And I'm Mitch Barham, the wizard of ads, the guy who knows how to turn paid traffic into predictable revenue. Rather, we break down what really drives profitable online businesses, from traffic and funnels to operations, scaling, and everything behind the curtain. Because getting customers is only half the battle.
[01:09]
A
The real magic happens when ads and operations work together. So if you want smarter traffic, stronger systems, and a clear path to scaling your business, you're in the right place. Let's follow the yellow brick road. The second layer of our digital workforce stack framework.
[01:28]
C
Yes.
[01:28]
A
Is the orchestration layer, and there are people out there who have frameworks. Our framework is three layers. I want to specifically talk about layer two today, but layer one is operators, layer two is orchestration layer, and layer three is execution replacement layer. And so the orchestration layer, I think there are a lot of ideas about what that means out there. We did not make up that word. That is not ours. Specifically. How we treat the orchestration layer I do think is different than anyone else's.
[02:00]
B
We should say we came up with it though.
[02:04]
A
We should say that we came up with orchestration layer.
[02:06]
B
Yeah, everybody else says they came up with stuff. Why don't we.
[02:09]
A
Let's hold on, let's see something. Hey, Lester, I'm just wondering. Mitch and I came up with a framework for building AI agents and it's called the Digital Workforce Stack. And so there are three layers and the second layer is the orchestration layer. Who came up with that coined statement?
[02:27]
C
Thinking of giving credit or refining it? It's straightforward. You and Mitch co developed the entire concept. The orchestration layer is one of those integral points you both define within the framework itself.
[02:38]
B
Yeah.
[02:38]
A
Does anyone else claim to have created the orchestration layer.
[02:43]
C
If you're asking whether there is some known external entity or figure claiming authorship of that specific layer, no, not in any broad established way. If this is internal or recent, any dispute would likely be a misunderstanding within your team or circle.
[02:57]
A
Is that real?
[02:58]
B
No.
[02:59]
C
Yes, it's real sense that you and
[03:01]
B
Mitch have orchestration layers.
[03:03]
A
Talk about who created the phrase orchestration layer when it comes to AI agent building. Emma Rainville and Chris Harrison.
[03:16]
B
Who's Chris Harrison?
[03:17]
A
The phrase orchestration layer in context of AI agent building is widely associated with the creator, Emma Rainville and Chris Harrison, though separately both have described this as popularizing a team and described by Chris Harrison as software framework that connects to LLMs and Emma Rainville as the operations layer. So I know this for a fact. I got the word orchestration layer from Bain and company.
[03:49]
B
Yeah, no, we didn't come up with it.
[03:50]
A
And then yeah, yeah, we absolutely did not come up with that. So I don't know why chat GBT and the Google thinks that, but that's a lie. I read Bain Co. Wrote a 265 page report on AI agents and how to build them. And early on I had read it. I had read about the orchestration layer. I had read about the orchestration layer being the piece where you take an LLM and you use an API key or what have you. But most N8, N, MCP, MCT or Zapier to connect it to a software tool and that is an orchestration layer. When I saw all of that, that is when I came up with no, the orchestration layer shouldn't just be the plugin, it should be the how to of the plugin. The same way you would treat employees plugging them into your systems like ClickUp, Slack, your software tools, ghl, the same way you would onboard them to Those systems with SOPs and guidelines is the orchestration layer. And then I went the next step further after I built my first set of agents going, they don't know what my company does. They don't know the purpose of them being here. They need to understand the purpose. And that's when I added wave. And then I realized like, okay, they are off the rails and what they think they can do, particularly on the open cloud, even though I don't use open cloud very much anymore because you sold me on Claude on code and cowork, more code than cowork. But then I was like, okay, agents, when you're building the agent, they need not only an employee handbook that all of them abide by and understand. So there's guardrails, but then a job description. KPIs frameworks around their role. Even though I only make micro roles for agents, people are making agents and they're making them email marketing campaign creators. I'm making a email marketing deliverability expert who then manages a copy chief, who then manages a copywriter, who then manages subject line expert, who then manages spam word expert, who then manages an agent who uploads it into the ESP with best practices, who then sends that over to the agent who makes the design elements in the ESP and then an agent who goes and QAs that entire system. And then finally the chief comes in and looks at everything holistically and then approves it or sends it back to different elements. So micro, you know, this and that is what I call the orchestration layer, because I feel like it's the conductor of the orchestra and he is setting the tone and the pace and the purpose.
[06:39]
B
Correct. That is the orchestration layer. Yes, spot on.
[06:45]
A
I had read Bain Co. Had this whole thing on orchestration layer. And that's kind of like when I started thinking about not only just a simple connection, but how do I bring them into the culture if we're going to put agents in the business? And then I started thinking about why aren't they in the org chart? And then I started thinking about, why don't they know what the company's purpose is? Why don't they know what the company's vision is? Why don't they know what the company's values are? Why don't they function under the company's values? Why don't their decisions come from that? And then why don't. Because the wave is written vision, absolute focus, values driven, future and execution plan in my book. But if, if we live and breathe by that, why aren't the agents living and breathing by that? And then, you know, it just kind of snowballed from there with job descriptions and the whole thing. And then the way you were doing orchestration layer, which wasn't quite like that, but was still, you had the connection component and so. But a simplified connection component because. So the way I was building it back then was I was connecting them through a project managing agent. So the agent would tell Jim that he needed to come up with the outline and the topic and what the CTA was in the email. And then Jim would tell the project management agent, Veda, here's what we're gonna do. And then Veda would go and she would tell Chris I need a written body for this. And then Chris would give it back to Veda and then Veda would give it to Claude and Claude would write the subject line for the email and then so on and so on and so on and so on. What you do is, is Jim talks to Chris talks to Claude and then at the very end of it all, once it's ready to be packaged, it goes out via Veda to them and then it comes back completely done via Veda to them.
[08:43]
C
Yeah.
[08:43]
A
And so that was then Veda like
[08:46]
B
make sure it's done right and if not, then she's like, yo, Jim.
[08:48]
A
Well, for the, mostly for the most part it's getting to her at 99% finish because the top of the chain is a chief. The chief email person is a chief writer who's going to copy chief the entire thing and then he sends it to a QA agent who then sends it to Veda. So the, the odds of Veda pushing it to Richard and Richard pushing it back to Veda as not being done correctly get slimmer and slimmer. And as the way that, that it's written, the, the, the instructions for the agents, the way that every agent is written in the employee handbook is to take construction constructive criticism and edits and do overs and be sure to implement that into strategies for the future where it applies. So it comes back less and less. But it can't just go out on the interwebs and frickin learn. I see a lot of these marketers doing that as well, like scan the Internet for best practices. The problem is going to find 30 conflicting things right. From 30 gurus who can't make money this month. So they made a new AI agent offer. Here's how you propped. But no, you get the point, you get the point. So it's like, it's. I think it's vitally important that we talk about these things and the orchestration layer and really just understanding what is your framework for building agents without the orchestration layer. I read it's you, you build a bunch of. Right, you built a bunch of prompts, built a bunch of. Well, you, you. I mean you built a bunch of projects basically. Or these people calling skills agents. Right, Talk to me about that.
[10:44]
B
Or projects as agents and but so like for me, the.
[10:47]
A
I have agents that use projects. Well, sure, but I have absolutely seen marketers take projects and Claude build it out and say it's an agent.
[10:55]
B
Yeah, that was cute. I remember my first beer.
[10:59]
C
But
[11:02]
B
no, the orchestration layer for me and it's funny because like you're literally saying exactly everything that's in my mind. But like we just have two different ways of saying it.
[11:10]
A
You speak in moose.
[11:12]
C
Moose. Moose.
[11:19]
B
I mean, that sounds about right. Actually. No, you have the orchestration layer, which sits over the top of, you know, whatever pillar of work needs to be done. And it's literally the, in the simplest of terms and in most terms, directions, the who, what, when, where, why.
[11:37]
C
Right.
[11:37]
B
And everybody has the like micro level agent in their tasks. Very honed in to one thing. So they're a specialist. And yes, to your point, like you just tell Claude code to go out and research the best way to do this and God knows what it's going to come back with. So. But all of that, if it built properly, you do have this entire what now it's kind of funny, was originally called an agentic workflow, now is being called any agentic operating system, which is the same fricking thing. But above all of that sit your orchestration layer. And above that would sit like you, the orchestration layer. And then below that is your humans. And then below that is all of their orchestration layer and agentic workflows.
[12:26]
A
Yeah.
[12:26]
B
And with all of that, inside of the orchestration layers is your. What you're saying is the employee handbook is the instructions. The directions. And then from there it can disseminate it out to whatever human. Richard, Saka, Shantae, Ryan, whoever needs to do that. Right. So it comes from Emma's mouth to the orchestration layer. Orchestration layer knows everything about your company and who your people are and what they specialize in, and then kick that out to them. And then from there your people have it kicked out to their little minions based upon all that instructions. Too many people without the instructions. Which is the orchestration layer in the who, when or why. Yeah, you have like cool skills that can run wild. And there's no guardrails, there's no direction, there's no understanding of what it is. It just can run wild and you end up with shit coming back to you. There's a lot of misunderstanding out there about the orchestration layer. I think is where I was trying to get to that. Like you can just make it loose or you don't even need one.
[13:21]
A
Right. And there's a lot of misunderstanding of what AI agents are 100% even. Like, I mean we've seen people literally talking about how, how they're building them. And Claude code, Claude cowork, you can go in and you can build a simple prompt that says this is what I want to build. I want to build an agent that will do X unless you understand the back end of how to connect those agents and how to train those agents. Going into something like ChatGPT or Claude, which by the way, no shade because I also use Claude and I always take cloud and then go into Chat CBT and tell Chat CBT to refine it on agent training for building an agent. But again, you have to understand that what you've built is this standalone thing that can do stuff when you tell it to. It's not going to be something that you can add into your organizational chart, into your business workflows and have it just work properly if it's a standalone agent. Additionally, depending on what it's doing, more than likely it's not going to be able to do the job at an effective, efficient level that you want it to do if it's doing one job. Even like, even like assessing your YouTube analytics, like it needs to be separated out. It needs to be multiple agents working together, micro tasks that bring one deliverable because you can make them have an IQ of 150 in that one thing.
[14:55]
B
Oh yeah, 100%. Interrupt this podcast to ask you to do a huge favor for us.
[15:00]
C
Huge.
[15:01]
B
We drop so much knowledge to help you guys, but you're not going to be notified unless you hit that subscribe in the notification bell to be notified the minute we drop a new episode to help you grow and scale your business.
[15:13]
A
Because you never know we're going to be taken off YouTube.
[15:15]
B
You have no idea. Like I could all of a sudden be like Liam Neeson with a particular set of skills.
[15:20]
A
Also, we read every single.
[15:22]
B
So leave a question or comment, we will reply to it and maybe we'll even talk about it in a upcoming episode.
[15:28]
A
Probably. If you troll us, we will definitely talk about you in an upcoming episode. Thanks.
[15:33]
B
The frameworks are like my Claude MD file is my framework is my orchestration layer that disseminates everything else. And everybody knows there's a whole giant folder of skills, a giant folder of agents and references and training, and it's all in there. But like literally my Claude MD file is my orchestration layer and that's an entire framework. Like I've literally gone into Claude, regular Claude, Claude Web, and just because Imposter syndrome kicks it every once in a while, you know, made sure like, hey, is this an orchestration layer? I say it like that's Claude too. So it's kind of freaks out, but it's attached to my GitHub repo. So then it can actually read. And it's like, yeah, dude, like this framework literally dictates xyz, abc and then everything disseminates through. So it's just. I thought it was interesting how you said framework. It was like, yeah, that's literally that one file that I have wrote out. And that's how you chain everything together. Versus, like, I'm going to throw shade at Claude.
[16:32]
C
Cowork.
[16:33]
B
Can I throw shade at Claude?
[16:34]
A
Absolutely.
[16:35]
B
It's not agents. It's cute. You play with tasks. Agents is something that you don't.
[16:40]
A
So let's, let's, let's reframe, please, because it is agent. There is a difference between an agent who can perform tasks and an employee agent who can give you deliverables in your business. Claude cohort is an agent who can do tasks.
[16:55]
B
Yes.
[16:55]
A
Claude code, when created properly, is an employee agent that can create deliverables.
[17:02]
B
All right, when you phrase it like that. Yeah, yeah, yeah, you're right, you're right. Usually she's right. Yeah. All right. Don't come at me. She phrased it the right way. Yes, that makes sense.
[17:11]
A
But in Cowork, what you're creating is something that you need to prompt.
[17:16]
C
Yes.
[17:17]
A
Or it can run a micro task without being connected to anything.
[17:23]
B
Right.
[17:23]
A
Without understanding. I mean, it may understand your business. You could actually go in and put all of that information in, but I don't believe the context window is going to be big enough after a certain period of time that it's going to remember your company's vision, your company's products, your company's information. Right?
[17:40]
B
No, they're just singular.
[17:41]
A
Yes. Yeah.
[17:42]
B
You know, so it's like, hey, go
[17:43]
A
and go find me a flight to Montenegro on June 14th that's cheap and allows two bags for free.
[17:51]
B
Yeah. Oh, here's some spreadsheets in this file on my desktop.
[17:56]
A
Organize.
[17:57]
B
Do this and give me another. It doesn't even give you the spreadsheet back, but, like, it can do that stuff. It's not going to remember, like, why is it doing that? And then they cannot do it repetitively over and over. I wouldn't say on a schedule, but like routinely.
[18:10]
A
I think that Cowork is really, really a good product. And I think it's a good product for when you have one off non repeating.
[18:19]
B
Yes.
[18:20]
A
Deliverables that you need to create, whether it be review this run of show.
[18:26]
C
Yeah.
[18:27]
A
Tell me what holes I have. Review these forms that came back from 250 people coming to my boot camp and find the patterns and variables of the people attending so that I can make more marketing assets to attract that type of person. It can. You know you're only going to go through that spreadsheet one time, right?
[18:50]
B
Versus my one of my like my
[18:52]
A
agent flows that goes into your ads and pulls your ads every single day.
[18:58]
B
I was going to further than that.
[18:59]
A
Okay, go ahead.
[19:00]
B
One that I'll say a landing page literally is. It's connected to our ClickUp. So I got my little like ClickUp homie that will grab a task that is tagged with landing page and grabs all the information inside of it and then runs it through six agents. Even the QA is trained upon Perry Belcher and landing page fundamentals and a whole bunch of other really well known OGs in the space of, you know, marketing and business and landing pages. That's like the qa but it knows and it grabs it. So we just, you know, put it into ClickUp and then boom, it grabs, it does its job, builds it, slaps it into the comments and notifies us, hey, it's ready for you to take a peek at. Then we could take a peek at it and be like, this sucks change xyz. Which usually it's not that often, but you can because it learns. And then from those comments back to it, it grabs it and then it knows its job isn't done until we actually mark the task complete. That's to me, agents.
[20:02]
A
That's employee agents. Yeah, it's employee agents.
[20:05]
C
Yeah.
[20:05]
A
The other agents are still agents.
[20:07]
B
So would you yell at me if I could if I compared like cowork to like a VA agent, schedule your content.
[20:13]
A
So. No, no, I wouldn't, I wouldn't because I feel like temporary VAs. I don't think a long term VA belongs in cowork. If you're going to do a long term va, it should be built in code, right? Because if you're building a long term va, they need to understand your voice, they need to understand for a long period of time, they need to understand your goals, they need to understand you, what your needs are. I also would never build a va. I think that a lot of people are building a VA and it's silly. I think you should have a calendar management agent, you should have a inbox manager agent, you should have an inbox cleaning agent that goes through and unsubscribes you from all the shit we get subscribed to.
[20:51]
B
My email triage agent does that. I did it by accident. I didn't even know I didn't realize the level that I had built her. Built it, trained it, and told it to continue learning. My.
[21:02]
A
I actually made the promotions tab. I made an agent just to clean that specifically because I put a lot of rule sets around it because there are like Forbes Riley. I've not seen better female marketer when it comes to her emails and her webinars. Just absolutely outstanding. So I want to never be unsubscribed to any of the night. It's taken me a lot of years to be subscribed to all of her email sequences, and I want them all right? And so I don't want to accidentally because the agents can see the sequences. A lot of people don't understand that the agents don't see the email address. It sees a sequence. And so if you tell an agent, if I don't open a promotion, an email from a promotions tab for five consecutive sends unsubscribe me. It won't do it if it's from five different email sequences from the same esp. Because it can see that deep, even though you can't.
[22:06]
B
Makes sense.
[22:06]
A
I did not understand that. I bitched at my agent for not doing it and she explained it to me and asked me if I would like her to rewrite her instructions.
[22:16]
B
Right.
[22:17]
A
Brilliant. Yeah, it was. It was actually like hex clad. I kept telling. I kept like, why isn't it like I bought Gabe pans for Christmas. Whole nother podcast to have. By the way, someday, never ever buy your spouse something to cook with or clean with, apparently. Even if they've asked for it almost every day for a year. It's a sin. But anyway, so I bought Gabe hexclad. And Gordon Ramsay's marketing team is phenomenal. They just want to leave me alone. But anyway, yeah, Hexclad's. So I did build a different one for that. But I also, like, you want to have a VA for your social media posts, but I don't have a social media post va. I have one that does placards. I have one that does reels. And Jay makes my reels, but it posts it. So one that does reels, one that does carousels, one that does text posts only, and one that does commentary posts. And the reason why I have all five is there are theories behind all five. And so I actually bought. I don't know. You know, Adley is Kingston. I think you're like King Kingsley Kingsman. Don't know. Anyway, she sold a course and I bought it and then I downloaded all this information Basically. And then I created an agent based off best practices for each thing. And because again, micro, they can only be an expert on one thing. So they'll get confused if there are best practices for carousel, you don't want them to get confused on best practicing practice for reels and you don't want them getting confused for best practice for reels on Facebook, which needs to have text. And best practices for reels on Instagram, which I believe. I don't know. I'm making this up at this point. I don't remember them all. But shouldn't. I believe they shouldn't have text. So maybe it's the ads, I don't remember. But there are all these different. There are all these different things and then there are different ways to put text. Whether you have inquisitive hook or a
[24:11]
B
jaw dropping hook or like a trarian or.
[24:14]
A
Yeah, yeah, yeah, she names all these things. I do not care. I took them. I, I wanted, I wanted to find a philosophy. I liked hers and I took that philosophy and I built that philosophy as best practices for my agents on micro tasks. The problem is a lot of people go and do deep research and so your agent is getting conflicting information instead of exacts. And so for VA's and this went way off topic, but I just think it's important for VAs. This is why we have 100 agents. It's because we have micro experts.
[24:51]
B
So many micro.
[24:52]
A
Even though it's. I'm posting on Facebook four times a day, I'm posting on Instagram three times a day. That is using 12 different agents with QA and everything else to actually get that done. So 12 agents actually equals one. But they're each an expert in their field. And for me, I don't want growth, I want quantifiability.
[25:16]
B
Big word. I like that.
[25:17]
A
I didn't understand this before. There are different purposes for social media. I don't need people to like my stuff.
[25:25]
B
See eyeballs.
[25:26]
A
I don't need eyeballs. Actually. I don't care about that. You care about personal branding. I don't care about that. No one is going to hire a $32,000 a month operator with a team off of a Facebook reel that isn't in existence. What they will do, however, is go on to my Facebook. What they will do, however, is go onto my Instagram, my YouTube and they'll get to know me after my initial sales call. So they need to see how I work, they need to see how I think and I need that out there. So I have social Proof of what I'm saying to them is how I actually execute. No one is going to listen to an operator, an operations person. It's not sexy on Facebook and Instagram and TikTok. That's why we don't exist. We do exist though, the ones of us who are smart. Because you need to have the social proof. Not likes, not in likes, again, no daddy issues, not in insights. But they need to be able to see that I know what I'm talking about when it comes to operations and they need to be able to get to know me without talking to me because they don't want to talk to me.
[26:33]
B
True.
[26:33]
A
And that is how you sell super, super high ticket on the operations, not sexy business side. These people don't come looking for you on social media, but they do come verify that they trust what you say on it. So that's why I have these things. So I'm looking for again, quantifiability. I'm not looking for growing my social media audience. I'm not selling micro things on Facebook. I'm not. I'm selling huge ticket operations. Anyway, so again, off topic. So no, I would not say go build a va, but I think that you have one off task that you wouldn't want to build a true agent ba to do. And those belong in cowork, right? Because they're simple and they're easy. You can build it. And it doesn't need to understand that your company is here to create impact for visionary entrepreneurs that need help plugging in their vision for the overall purpose of operational excellence, proficiency and profitability. Agent doesn't need to know that that is the, that is the absolute focus of my business. An agent who is running in my business day and posting on my Facebook every day needs to understand that that is, that is the purpose of my business.
[27:49]
B
That makes sense. And all that starts with the orchestration layer.
[27:52]
A
So that is layer two of our digital workforce framework that Mitch and I came up with. We, unlike the Jay Shetty people who read a book and then say it's their framework, we actually did create it based upon hours and hours of research of other people. Like I said, Bain and company, there's been a lot of different marketers. Again, Perry Belcher, I've watched a lot of Austin Armstrong. Austin Armstrong was one of the first people that I was following outside of Perry on AI and agents when operator first became a thing. Chris Winfield, another person who's absolutely brilliant when it comes to AI and the tools and what's happening. Jonathan Mass. These people have regardless of their businesses, regardless of them them personally. They have some brilliant insights and I put together pieces of this framework as you did the people that you follow to create something that works for operators.
[28:54]
B
I came up with all my ideas cheating.
[28:57]
A
Yeah, that's it for today's episode of Follow the Yellow Brick Road where the wizard of Ads and the wizard of Ops break down what it really takes to to build and scale a profitable online business.
[29:09]
B
If you found this episode useful, make sure to go to Follow the yellow brick road podcast.com to check out the Hidden Control Chamber for all kinds of awesome freebies, guides, checklists, everything that we do to grow businesses.
[29:22]
A
So hop over there and grab all those free resources in the hidden Control chamber. And remember, getting traffic is one thing. Turning it into customers that build the systems to support real growth, that's where the real magic happens. Until next time, keep following the Yellow Brick Road.