Building an AI Operations Team with Rachel Woods

Danyel McLellan [00:00:01]: 

Welcome to the Agency Management Institute community, where you’ll learn how to grow and scale your business, attract and retain the best talent, make more money, and keep more of the money you make. The Build a Better Agency podcast presented by White Label IQ is packed with insights on how small to midsize agencies are getting things done. Bringing his 25 years of experience as both an agency owner and agency consultant, please welcome your host, Drew McClellan. 

 

Drew McLellan [00:00:31]: 

Hey everybody, Drew Mclellan here from Agency Management Institute, bringing you another episode of Build a Better Agency. This week we are live at MACon 2026 in Cleveland, Ohio. So we are hanging out with thousands and thousands of people thinking about and talking about AI. And I have a special guest that I would not have found anywhere but here at MACon. So I’m super excited about that. And I’ll Tell you a little bit more about her in a minute. Before we get started, of course, I want to say thank you to our friends and our supporting sponsor, White Label. As all of you know, they’ve been a sponsor for many years, and what they do is they come alongside agencies and they are your white label design, dev, PPC, AI automation builds. They’re building out CRM systems. So if you just have a lot of work coming in and your tech team can’t support it all, They’re a great resource. If you don’t have a tech team and a client needs a website or a landing page or some AI automations, they are your go-to. And the beautiful thing about White Label is they were born from an agency. So they understand your business model. They understand what it looks like to partner with an agency and how to make sure that they give your client the amazing experience you want them to have and still leave you a buck or two in the process. So head over to whitelabeliq.com/ Am I to learn more about them? All right. So quick question for all of you. Who in your agency owns AI? Not who’s good at it, not who built the custom GPT that everyone uses for meeting notes, but who really owns it? Whose job description is it in and who gets reviewed for it? If you just went blank, you are in good company. Most agencies that we talk to have bought the tools. And they have one or two people who are really great with them, but they have not seen it show up in their margin. It is not woven through their team and their processes, and they’re not really sure how to make it just part of their daily life at the agency. So my guest today would say that is not a tools problem. It’s an org chart problem. So Rachel Wood started out as an AI data scientist on Facebook, on the Facebook ads research team. Then she built and sold a startup on GPT before ChatGPT was a household name. After that, she ran an AI operations agency that only worked with marketing, PR, and professional service firms. So she has literally owned a shop like yours. She’s been inside shops like yours. And when she does, she’s helping rebuild real workflows. Today, she is CEO and co-founder of AI Momentum Protocols, or AMP. And she runs a community of more than 1,500 AI operators. We’re gonna talk about who in your shop should own AI, what goes into that role, and where those freed up hours can actually show up on your P&L. Rachel, welcome to the podcast. 

 

Rachel Woods [00:03:40]: 

Thank you for having me. I’ve been looking forward to this. So excited. 

 

Drew McLellan [00:03:43]: 

Me too. So you’ve been in a lot of agencies of all sizes and shapes. When an agency owner says to you, We’re using AI, but honestly, it’s not really changing how we function. What’s your diagnosis from that sentence? 

 

Rachel Woods [00:03:58]: 

So, the first thing that I ask is, okay, well, what use cases have been successful for you so far? And from that, I can either diagnose one of two problems. Either one, that across their team, they’re kind of still in the basic, well, we’re using it for emails, you know, we’re tinkering here and there with different things, but they don’t have a lot of good use cases. Or the other that actually is more common, and usually is the people that, that feel like AI should have done something by now, but it hasn’t. And they’re like, oh, well, we actually have an AI-powered system that powers blah, blah, blah, blah, blah, part of our service, or gives our customers in, you know, in-depth reports. Like they have something that’s a use case that feels really powerful, but then they still feel like they have this gap. And if you’re in that second position, the problem is that you are looking at AI as something like a project, like an initiative. 

 

Drew McLellan [00:04:48]: 

Yeah, right. 

 

Rachel Woods [00:04:48]: 

Versus a way of operating. 

 

Drew McLellan [00:04:50]: 

Right. Yeah. It’s, it’s sort of like before computer, after computer in agencies, right? 

 

Rachel Woods [00:04:56]: 

Exactly. Or before buying that fancy software. 

 

Drew McLellan [00:04:59]: 

Yeah. 

 

Rachel Woods [00:04:59]: 

The after buying the fancy software. Right. And how many times in the past have you bought fancy software that you thought was gonna completely change your margins or how you run your business and it didn’t? 

 

Drew McLellan [00:05:08]: 

Right. Every project management software on the planet. I think people felt like— 

 

Rachel Woods [00:05:11]: 

We won’t name names, but— 

 

Drew McLellan [00:05:12]: 

That’s right. No, no. Right. So in your work, you sort of identified 2 things, tool chasing and settling. So, talk a little bit about those 2 problems. 

 

Rachel Woods [00:05:22]: 

Yeah, so I think the one that’s easier to recognize is the settling, right? Where, uh, it almost feels like you’ve checked a box. And one thing that’s really common is whenever, um, you know, top-down as a company, you wanna be able to say you’re using AI. So, what do you do? You probably put it maybe in a, a mandate, if that’s something that you do as a company, or maybe put it in your quarterly rocks, right? If you’re an EOS or, But you’re like, we wanna use AI. So then what does everybody on the team do? They try to demonstrate that they’re using AI for different things. And again, you kind of hit this state where it’s like, yeah, we’re all using, you know, ChatGPT or Claude, and we, we feel like we’re, um, we’re there, right? And so that’s like this point of settling. But then again, you know that there is probably more, you just feel stuck of like how to unlock that. The other common thing though, again, is this like tool chasing, and this can happen one of two ways. One is looking for either your own internal projects that you are trying to build out custom AI tools for, kind of like we described, um, or buying custom AI tools from vendors, right? Of here’s this in-depth AI-powered, you know, ad creative reporting mega system, and then buying that. But this also happens with tool chasing the latest agent. Right? The GrokBot, the Muse, the this, the that. And I think this is especially common when you’re the owner who is interested in AI, that you turn around, you know, every Monday it feels like you have like, hey team, we gotta be using this, right? And that’s actually, in a lot of cases, worse for a team than being in the settling. ‘Cause at least in the settling, everybody feels like they have ground to stand on. You can build on that stable ground. The tool chasing just feels like whiplash. Yeah. And so you’re really not gonna get progress other than like cool demos. 

 

Drew McLellan [00:07:09]: 

Well, and what we’re seeing is in a lot of agencies, there’s one or two people. One of them is almost always the owner, and then somebody else who is a little further along. They’ve built out some processes or something like that. And then everybody else is just using ChatGPT or Claude or whatever it may be, kind of on their own independently. So you’ve got all these different people using these tools. In like a little vacuum, but there’s no SOP across the organization and there’s no sort of mass lift in terms of efficiency or effectiveness or whatever it is you’re shooting for because they’re not sharing the same process, tools, outcome sort of methodology. Are you seeing that too? 

 

Rachel Woods [00:07:56]: 

Yeah. And that’s something that we call Really going from being just using AI to actually operating with it. 

 

Drew McLellan [00:08:04]: 

Right. Yeah. 

 

Rachel Woods [00:08:05]: 

And I’ll say that maybe the, the warning up front is probably a lot of the stuff we talk about today is not gonna feel super sexy, ’cause at the base it’s really just operations fundamentals, 

 

Drew McLellan [00:08:15]: 

right? 

 

Rachel Woods [00:08:16]: 

Built in this new way, um, of being, you know, AI-powered operations or what we call AI operations. But that’s the gap, right? And it’s, it’s no different than like hiring, you know, uh, smart people on a team, but having no shared way of working together or having a bunch of really, you know, powerful tools, but they’re not connected. The same thing is probably happening in your AI systems and that’s what’s holding you back. 

 

Drew McLellan [00:08:37]: 

And the consequence of that is, I, I, I think, I think the agencies that have adopted it at the owner level and maybe some, some at the, at the rank and file level, they’re afraid they’re gonna fall behind. And so, they’re doing anything and everything they can to play with it, to learn it, to use it, but it’s very isolated. 

 

Rachel Woods [00:08:59]: 

Right. 

 

Drew McLellan [00:09:00]: 

Yeah. 

 

Rachel Woods [00:09:00]: 

And it feels like every new project or new use case, again, there’s that word project again. 

 

Drew McLellan [00:09:06]: 

Yeah. 

 

Rachel Woods [00:09:06]: 

Like, it feels like you’re reinventing the wheel each time. 

 

Drew McLellan [00:09:09]: 

Right. 

 

Rachel Woods [00:09:10]: 

Like, you make something cool, then it’s like, okay, now how do we get the team to use this one? Right? And again, if any of that, if it feels one-off of how you’re approaching it, then the answer is you have to kind of go back down to the basics of figuring out what is your operating system? It’s like the DNA of the business. How do you get work done? 

 

Drew McLellan [00:09:29]: 

Right. 

 

Rachel Woods [00:09:30]: 

And is AI truly integrated into that? 

 

Drew McLellan [00:09:32]: 

Into all the workflow. 

 

Rachel Woods [00:09:33]: 

Yeah. 

 

Drew McLellan [00:09:34]: 

Yeah. 

 

Rachel Woods [00:09:34]: 

And what’s good news, especially if you’re somebody that in your team that does have those 1 or 2 people that are pushing ahead, great. You have the first piece of the talent that you need in order to start doing this. But you have to make sure that, um, and I call this oftentimes Your team is getting stuck in the, this is a nerdy term, but the bimodal distribution. 

 

Drew McLellan [00:09:56]: 

I think everybody talks about the bimodal distribution. Yeah. Right. 

 

Rachel Woods [00:10:00]: 

But in case you’re not seeing this, you know, I’m drawing a way to like 2 humps, right? You know, you can imagine it like you run, uh, you know, you run an A/B test and the data actually looks like A was better than B. And you look at the 2 graphs and you see 2 humps, like that same thing is happening to your team. And the issue is that. the hump or the group of people that are ahead, they need to either turn around and actually help the people behind, or you need to decide what you’re gonna do with those people. 

 

Drew McLellan [00:10:29]: 

Right. 

 

Rachel Woods [00:10:30]: 

That’s the reality. 

 

Drew McLellan [00:10:31]: 

Right. Is there, can you envision a world a year from now, 5 years from now, whatever, whatever it is, it’s AI time. It’s like dog years, right? Everything goes faster. Can you envision a world where an agency can afford To have AI naysayers on the team? 

 

Rachel Woods [00:10:48]: 

I think, I mean, I would even question whether you can afford that now. 

 

Drew McLellan [00:10:53]: 

Agreed. Right. Yeah, I think that’s one of the challenges for a lot of agencies. All right, so let’s, let’s talk a little bit about the org chart then, ’cause part of what we’re talking about is, you know, you’ve got these isolated individuals who are invested or making the tools or playing with the tools, but it’s happening in these bubbles. So, You have sort of mapped out 5 key roles that every agency has to have to really operationalize AI and to make it something that is just naturally woven like oxygen through the workflow. So walk us through those 5 roles. 

 

Rachel Woods [00:11:27]: 

And before we get into the roles, I’ll also just acknowledge if anybody’s listening to this and going, holy crap, no way I have room for 5 new headcount. Don’t worry. This can be multiple hats worn by one person. 

 

Drew McLellan [00:11:41]: 

Right. We’re not saying hire these people, right? Yeah. We’re just saying— 

 

Rachel Woods [00:11:43]: 

this can also be multiple. So like one person in your organization can wear multiple of these hats. So maybe let’s call these like the 5 hats, right? So the 5 hats that you need is first, you need to really know what your job is as the owner, 

 

Drew McLellan [00:11:56]: 

right? 

 

Rachel Woods [00:11:57]: 

Which is that your job, I hate to break it to you, is not to build all the solutions and all the systems. 

 

Drew McLellan [00:12:03]: 

Which is totally what’s happening right now, right? 

 

Rachel Woods [00:12:06]: 

You can if you wanna find somebody else to lead company. Run the business, right? Run the business. 

 

Drew McLellan [00:12:10]: 

Right. 

 

Rachel Woods [00:12:10]: 

And work on the business, not in it. 

 

Drew McLellan [00:12:12]: 

Right. 

 

Rachel Woods [00:12:13]: 

But building this, building the systems is working in the business. So, you’ve demoted yourself by putting yourself in that position. 

 

Drew McLellan [00:12:20]: 

I know a lot of people getting a little tear in their eye right now, ’cause they’re loving playing with these new tools, right? 

 

Rachel Woods [00:12:25]: 

And you know what? That’s so important because what you need to do is you need to be the AI visionary, right? So, you need to know enough about what’s possible, and you need to know where you want to, guide the business in order to set a clear vision. And when I think about vision in this case, it’s really about making sure that the team has the right priorities and are, uh, using AI in the right order for the right problems that are actually gonna make a business impact. Like, that’s your job. So if your team is going off and using AI for like random stuff, you’re like, that’s cool, but that moved the needle nowhere. That’s your responsibility to be the one really setting the vision for it. 

 

Drew McLellan [00:13:03]: 

And probably also the governance, right? Like, Here’s, here’s our boundaries of when and where we use AI. Here’s how we talk about it with clients. Here’s where it shows up in the MSA. Again, to your point, not a lot of sexy things, but a lot of important things. 

 

Rachel Woods [00:13:16]: 

Yeah. 

 

Drew McLellan [00:13:17]: 

Yeah. 

 

Rachel Woods [00:13:17]: 

And then the other responsibility is that you need to be the prime champion. Yeah. You have to be, and this is not the champion of AI. This is the champion of the team’s success moving into operating in this way, right? So highlighting the momentum. But that’s really your job. Then the next role that you need is really what we call the AI operator, which is the person that’s really sitting at the core of taking the business processes and what, what things actually need to get done and translating those into things that AI is doing. The way that we think about this is, at its core, what is a business? Well, it’s a bunch of systems and process all working together to create reliable outcomes that customers pay for, right? 

 

Drew McLellan [00:14:02]: 

Right. 

 

Rachel Woods [00:14:02]: 

And so, you need to have somebody that is translating the things that are, that, you know, are done in the business into now systems and processes that AI is doing, right? And that person is, again, we call it the AI operator, and their skillset can only, not only be applied to themselves, but also to other people across the company. We’ll get into a little bit more That they shouldn’t be the only one building, but they really are like the lead operationalizer of AI. Okay, so that’s the second role. The third role is an AI implementer. So we specifically separate out the operator and implementer when we look at teams for one reason, which is that if you have a really talented engineer, that engineer doesn’t necessarily mean they’re gonna be really talented at business operations and process. 

 

Drew McLellan [00:14:54]: 

Ah, okay. Makes sense. 

 

Rachel Woods [00:14:56]: 

It can’t, it can’t be the same person. 

 

Drew McLellan [00:14:58]: 

Yeah. 

 

Rachel Woods [00:14:58]: 

But especially if you’re a larger organization, I’d really encourage you to think about your current people that are building AI stuff. They might be more the AI implementer type, and you’ll be able to run a lot faster if you pair them up with somebody who is more of an AI operator. The background of an AI operator is typically like an operations person. I like to call it the ClickUp queen. 

 

Drew McLellan [00:15:17]: 

Yep. 

 

Rachel Woods [00:15:18]: 

You know, the Airtable king, right? The person who builds the systems and more importantly also rolls them out to the team and everything. But those are, those are 2 roles that again, larger teams can have separate. And that’s a way to, uh, go even faster with, uh, building out all of these systems and processes. Then if you’re like, okay, Rachel, this sounds like a lot of like infrastructure roles. What are we actually like building? 

 

Drew McLellan [00:15:39]: 

Right. 

 

Rachel Woods [00:15:39]: 

Well, what you need to do is you need to, again, look at the systems and process across your business, especially if you look at, um, you know, the core drivers of your agency, right? Getting new clients, onboarding those clients, and delivering for those clients, then what are all the processes inside those and who are the subject matter experts of those processes? And that’s the 4th role, which is who’s gonna be the person on the team who is really, um, the best. And if you’re like, man, if we could just clone Jared, 

 

Drew McLellan [00:16:07]: 

right? 

 

Rachel Woods [00:16:07]: 

Like every client would love their onboarding process if we could just clone Jared, ’cause Jared does the best audits or whatever the, you know, um, Jared’s role is. That person needs to work with the AI operator to basically take the process that Jared follows, or maybe even define the process. 

 

Drew McLellan [00:16:24]: 

Right. Because a lot of times it’s done instinctually and it’s not written down. Yeah. Right. 

 

Rachel Woods [00:16:28]: 

Yeah. Which is where it’s helpful to have, again, that kind of back and forth because a good AI operator will sit there and be like, okay, I don’t really understand why you’re making this judgment call. Can we talk through that? You know, and they’ll be able to identify where Jared still needs to be in the loop, you know, those kinds of decisions. But you need to figure out who’s gonna be your subject matter experts across the team. And that’s how you identify the different processes again that you’re gonna set up, uh, with AI. And then the 5th role is everybody else, right? Because this, this is not a— 

 

Drew McLellan [00:16:54]: 

Everyone who’s gonna use it. 

 

Rachel Woods [00:16:54]: 

Yeah. Yeah. The end user. And one of the common misconceptions that we see is people think that you can get away with just a couple people on your team, like knowing how to do this. When I define these, um, these 5 hats, we actually do this like for each project or for each initiative, meaning that then there are What you really want is for everybody on the team to be able to do this in a micro way for themselves. So, I’ll give you a specific example. Everyone should have at least like one personal playbook that they made that’s just for something that they do that helps them speed up their day, where they have the experience of taking a process or making a process and getting AI to do that process for them. And then in an ideal world, that now creates value for them. They might share it across the team, um, and that becomes something bigger. But every single person needs to have the experience of doing this for something in their work at At least one. 

 

Drew McLellan [00:17:47]: 

Every single person needs to, for themselves, fill all 5 roles on their personal playbook or project. 

 

Rachel Woods [00:17:54]: 

Exactly. Yeah. We train the teams that we work with on these being, you know, very basic small things, right? But for example, a common, uh, trope we see is if you have a meeting recap GPT, right? Ask your team how much people are editing the recaps that come out of it. If you have a Frankly, if you have a good team, oftentimes each person is still editing and adding their little flair. 

 

Danielle McCier [00:18:19]: 

Right. 

 

Rachel Woods [00:18:19]: 

Great. That’s a perfect personal playbook because the way that, even though we have the same kind of meeting recap, like structure as a company, uh, the way that you want your meeting recap to sound like you versus the meeting recap to sound like me is going to be a little bit different. And so those would be great personal playbooks to do. So again, then I have, I’m more bought into this idea and I have the experience of getting AI to help me. And driving the process myself, that’s gonna make it easier for me to work within whatever other, uh, projects and processes get built by the overall team. 

 

Drew McLellan [00:18:48]: 

Sure. I’ve seen the process, I’ve done it, right? And I, and I understand the roles. So if you’re the operator and I’m the SME for this particular project, I get what you need from me for you to do your job so I can do my job. 

 

Rachel Woods [00:18:59]: 

Exactly. 

 

Drew McLellan [00:19:00]: 

Yeah. Yeah. Yeah. Okay. So in a typical agency, you were saying the operators are typically like, like the ClickUp queen or somebody who’s on the operation side or. project management, traffic side. In a 25-person agency, typically where do these roles live? 

 

Rachel Woods [00:19:15]: 

So at that size, we see it in one of two places. Either this is like, sometimes at that size, I’d be curious, you know, your, um, input on this, but you see the ops team start to almost kind of feel like it’s gonna be this separate little entity, right? Where you have someone that becomes more cross-functional, and that could either be kind of like a mid-level person that’s more cross-functional. That’s a great, like, key AI operator. Or sometimes it can be, you know, your managing director or your COO or someone that is, uh, higher up on the leadership level, but still just the size of the business. 

 

Drew McLellan [00:19:55]: 

Director of ops, COO, whoever that may be, the head of the project management department, someone in that role. Okay. So visionary’s almost always the agency owner. What if it’s not? What if the agency owner is AI anxious? 

 

Rachel Woods [00:20:09]: 

So that’s where you need to surround yourself with people that complement your weaknesses and strengths, right? And so we do see sometimes, you know, maybe it’s a president, maybe it’s a managing director, but again, you have to have somebody that’s not the primary AI operator that’s sitting in that visionary seat because otherwise You just get stuck too much in the weeds and you don’t actually work on the right things. And that’s probably what’s happening if you haven’t separated out these hats yet. 

 

Drew McLellan [00:20:38]: 

Okay. So visionary agency owner or somebody at that sort of C-suite level, the, uh, operator is gonna be somebody on the ops side. And then what about the implementer? 

 

Rachel Woods [00:20:50]: 

So the implementer, if you have a tech team, that’s probably who’s already kind of building your custom AI tools internally. 

 

Drew McLellan [00:20:57]: 

Right. What if you don’t? 

 

Rachel Woods [00:20:58]: 

Well, one of the things that I, I find funny, I still have yet to meet, uh, an AI implementer engineer that was like mad that they got to focus only on implementing instead of having to also do the AI operations piece, right? They’re happy because, because their skillset, and I even have this dynamic with my sister who’s my co-founder, she’s a brilliant implementer and she’s so much happier when she’s wearing that hat in projects and they go faster and they’re better, right? Um, where I’m wearing more the operator or visionary hat depending on the thing. But if you don’t have that talent in-house, another good news about separating out these roles in this way is that a good AI operator, you could position that person to work with contractors as implementers pretty easily because it’s easy hourly work building automations, testing automations, fixing things, maintaining them. It’s very clearly scoped if you do it right. 

 

Drew McLellan [00:21:50]: 

Yeah, I mean, that’s an example of that. In the intro, I was talking about what our sponsor does, White Label. They’re doing a lot. They’re doing, they’re playing that role for a lot of agencies who have no tech talent in-house and need that partner. So, okay, so that you could carve that out if you needed to. 

 

Rachel Woods [00:22:05]: 

And then one other thing that’s helpful, I think on this, that you might think, oh, well, can we also do that for the AI operator role? Right? Could we also have kind of like an outside contractor or, uh, 

 

Drew McLellan [00:22:17]: 

AI consultant, 

 

Rachel Woods [00:22:18]: 

consultant do that? And what I recommend is if you. You know, truly are trying to like dip your toes in, it’s not a terrible way to start because it will help you start getting momentum. But just like having, you know, a couple hour a week ops manager, like you feel that bottleneck so fast. And so it’s one of the highest leverage roles. 

 

Drew McLellan [00:22:41]: 

And they’re not available in the hours you want or the day you want. Yeah. 

 

Rachel Woods [00:22:45]: 

And they don’t really have the deep core belief of an understanding of the business. Part of the operator’s responsibility is making sure that the things that get set up are actually used and the team is trained and rolled out and that you’re iterating on the feedback and things are getting better and not abandoned, just sitting on the shelf. And that’s really hard to do if you are in a small scope. 

 

Drew McLellan [00:23:03]: 

Not embedded, right? Yeah. Yeah. So it’s interesting, and I wanna get to the other roles, but you think about, you know, you’ve built a company that helps agencies build this out and you could have built a company that just did it for agencies. So why didn’t you? 

 

Rachel Woods [00:23:18]: 

So in fact, that’s how we started, right? We would go into companies and we would, again, really like map out their operations, step in. And, you know, I’m, I’m sure people here can relate to the early days of an agency, right? I was the AI operator actually inside 39 businesses in a single year when we first started. Oh God. Yeah. Um, because of course I’m using AI to make myself efficient and stuff, but the, um, the reality is the model. Just doesn’t work in the way that, that it could. Again, we were good at getting people that initial momentum to understand, hey, here’s how these projects can go. But we moved to a, uh, training and enablement approach because if I’m sitting here with you, even if I’m spending 10 to 15 hours a week, again, I’m still not embedded enough to be able to really drive the operational change that this requires. You know, as, as again, as a part-time operator, you can do a couple of projects, but you’re gonna be limited, uh, in the scope that you can have. 

 

Drew McLellan [00:24:18]: 

Yeah, it’s interesting. I, I’m equating it in my brain to you and I are going to Italy. I have Google Translate on my phone, which is functional, but on a surface, you speak Italian. We’re gonna have a very different experience. And what you’re talking about is Hey agencies, you can bring someone in to do this for you, but if you do, it’s gonna be clunky and on the surface as opposed to just woven into the DNA of your agency. 

 

Rachel Woods [00:24:45]: 

And the other piece is, I think if you are sitting looking at these 2 options, you might be underestimating your current team. 

 

Drew McLellan [00:24:52]: 

Hmm. Right. 

 

Rachel Woods [00:24:54]: 

And especially if you have like a really badass operations person or a couple of them, or even your, you know, your directors of your different departments are people who’ve like made SOPs, trained people, you already have the hard part, 

 

Drew McLellan [00:25:09]: 

right? 

 

Rachel Woods [00:25:10]: 

Frankly, learning to take that skillset and now teach AI to do those processes instead of having humans do those by hand, that is just one extra tool in an existing tool belt that you probably already have on your team. It’s gonna be cheaper, faster, and a lower risk way to at least get to the next level before you start going and trying to hire and bring in outside talent. 

 

Drew McLellan [00:25:32]: 

And you’re gonna have better adoption internally because when we build something, we believe in it more than when somebody builds it and hands it to us, right? 

 

Rachel Woods [00:25:40]: 

Totally. Yeah. Yeah. 

 

Drew McLellan [00:25:41]: 

Yeah. Okay. So, uh, next role, the SMEs, those are probably scattered throughout the entire agency, right? 

 

Rachel Woods [00:25:48]: 

Correct. Yeah. 

 

Drew McLellan [00:25:49]: 

Okay. And then end users are everybody. 

 

Rachel Woods [00:25:51]: 

Yeah. 

 

Drew McLellan [00:25:51]: 

So really it’s visionary agency owner, or operator is somebody sort of at mid to high level on the ops side. An implementer is that technician, the builder, right? Okay. If I’m going to make somebody the operator, if somebody is listening to this and they’re like, oh, I know who the operator is in my organization, what goes, what goes into that job description? How do I help them understand what I’m asking them to do? And how do I measure their success? 

 

Rachel Woods [00:26:20]: 

So there’s 2 key frameworks that we teach that I obviously don’t have time to go into fully today, but I can give you the high level so that then you can start to think about how this maps to your people. So one is, uh, just like any ops role, there’s gonna be different levels of seniority in this, right? And so kind of the base level of an operator is that it’s somebody who can take a clear, uh, business problem needs to be solved and an idea already of what is the process or the specific piece of what we do that we’re trying to get AI to do. And then, um, we run what we call a craft cycle, which is really about First, getting a clear picture from the team, writing down those judgment calls, really decomposing everything, and making the decisions and trade-offs you’re gonna need to make in order to get AI to do it. So they run that process. They also need to figure out how to make this like a realistic thing, right? They might need to scope it down and manage, you know, project manage that. If they are, and that’s the R of CRAFT, if they are the one building, then A is AI-ify. If not, they’re working again with that implementer. Implementer, contractor, whoever that is, uh, to implement and keep that project on track. And then F is a feedback cycle. So this is where this person needs to be running it and testing it and— 

 

Drew McLellan [00:27:33]: 

Right. Oh, I forgot about this part. 

 

Rachel Woods [00:27:35]: 

And getting, uh, SME feedback, right? Of like, hey Jimmy, I, I just ran like the last 20 clients, uh, that we onboarded through the new onboarding playbook. Uh, you know, here’s the AI output. What do you think? Let’s go through this in detail. And then now I have the skillset to action to get AI, um, to be better at where it’s Uh, you know, uh, messing up. And then T is team rollout, which is actually making sure that this gets rolled out, trained, adopted. And that is the number one thing that people just go like, oh yeah, it’ll just, we’ll just announce it on a Zoom call one day. Exactly. I hate to break it to you, that’s not how real, you know, organizational change gets rolled out, right? You have to think about how to habit stack, how to make sure that people remember to use it, how to hold people accountable. What does success look like? And even in running that, um, that craft cycle, right, that’s, that’s how you can think about having, um, distinct processes across your organization that now should have whatever the success metric is, right, of faster, able to, you know, do more work in the same amount of time, uh, higher quality results, even, um, improved consistency. And we can go into a bunch of examples of like, what are some of these things that you can be Um, we call it playbooking, right, across your work. But that’s the core of what an operator, uh, should do. If you do have somebody who is more senior, they should also be responsible for identifying and prioritizing those use cases. Um, and we have a different framework that we call mapping, which is really about figuring out what’s the business breakthrough and what are all the ways that we could decompose that and then figure out how AI could either slot into how we exist, like how we do something in the existing way, or how do we reimagine how we do something so that then we have a new way that again drives that business breakthrough. But if you don’t have a senior person, you know, you’re gonna need to expect that you’re either gonna need to be more involved in that process as the visionary or train them up so that then they can be making those strategic decisions. 

 

Drew McLellan [00:29:31]: 

So as you’re describing the craft framework, I’m thinking, huh, I wonder how much It’s the person and how much a smart operator has actually created AI tools to help them ask those questions and discern what is all the parts and processes. ‘Cause I was gonna ask you like, how do they know what questions to ask? And I was like, oh, I bet they use AI to do that, right? 

 

Rachel Woods [00:29:55]: 

Yeah. So that’s one of the funny things, at least how we teach it, is, um, I’m a big believer and you need to do anything that you’re gonna be the expert for, or, uh, being, you know, operating a system of AI doing something, you need to do that thing by hand first, personally, so that you actually have the tactile experience doing it. But then, of course, you’re going to get AI to do it. And then, and so even in our program, we have like a, okay, now you have permission to playbook the craft cycle that we follow. So. 

 

Drew McLellan [00:30:25]: 

Yeah, ’cause that would be faster. 

 

Rachel Woods [00:30:27]: 

Totally. Yeah. 

 

Drew McLellan [00:30:27]: 

Yeah. 

 

Rachel Woods [00:30:28]: 

Yeah. 

 

Drew McLellan [00:30:28]: 

Okay. 

 

Rachel Woods [00:30:29]: 

You should be doing nothing by hand eventually, right? It’s just, it’s just this. Iterative process. Uh, and that’s, that’s why, again, if we go back to when you think about the project you’re really proud of, that’s your shining use case that hasn’t really made the impact you want. Imagine instead that you’re doing that on repeat, continuously unlocking more and more and more time by doing this, this cycle, this way of working across the entire business. That’s how it actually compounds and you actually start to unlock like the level of impact and, um, Bottom line that you want. 

 

Drew McLellan [00:31:01]: 

Yeah, it’s interesting. I, I saw a research report the other day that was saying, you know, business owners are saying AI hasn’t saved them any time or any money. And I think, well, to your point, they haven’t done it long enough or deep enough yet to sort of see the impact. So it’s sort of like saying, well, I tried to ride a bike once and I fell, so I can’t ride a bike, as opposed to I gotta keep at it and keep iterating. And like you said, compounding the value. Yeah. So that eventually it’s like, oh, now I see the equation, right? 

 

Rachel Woods [00:31:30]: 

Yeah. Think about it like you have the key to unlocking unlimited time. You just need to do it on a cycle. 

 

Drew McLellan [00:31:37]: 

Yeah. As somebody who loves to cram as much stuff into time as possible, that’s one of the things that I love about AI. 

 

Rachel Woods [00:31:44]: 

And maybe just while we’re on this, yeah, to really make it real, I have just a quick story. 

 

Drew McLellan [00:31:49]: 

Great. 

 

Rachel Woods [00:31:49]: 

That, um, is personal. So earlier this year I was like really sick one night. And I dunno if you can relate, but I’m sure a lot of, you know, owners, we all feel this way when you get sick. And I’m talking really sick and you’re not gonna be able to work the next day. Like you feel this existential dread, right? Especially if you’re a little bit still too involved in the business, cuz now you have your, your key driver of the business out. And so projects stall momentum. And your calendar’s full. Your calendar’s full. 

 

Drew McLellan [00:32:15]: 

Right. 

 

Rachel Woods [00:32:15]: 

You gotta reschedule all that stuff in an already packed calendar. Right. Your email inbox, and then you’re just hoping it’s gonna be just a one-day thing. Much less if it’s going to be multiple days. So this happened to me earlier this year. And again, I personally work this way of this, like stacking up these systems. And so even though normally getting sick, very stressful, this time all I had to do was open up my laptop and I typed one single word, which is todaying. That is the word that starts my morning routine playbook that then my AI agent goes and figures out all of its kind of baseline chief of staff work that needs to get done of email, calendar, all that stuff. But then also any other deliverables across the business using other processes and other playbooks in our marketing, in our sales, following up on leads. Like I could list out all this stuff, but from typing one single word, and I have my AI like have its receipts of how much time it saved, 

 

Drew McLellan [00:33:10]: 

right? 

 

Rachel Woods [00:33:10]: 

It did over 8 hours of work. 

 

Drew McLellan [00:33:12]: 

Nice. Yeah. 

 

Rachel Woods [00:33:14]: 

And that is entirely possible with how the tools work today and what you are able to do and your team is able to set up. It’s just a matter of committing to really unlocking this cycle. 

 

Drew McLellan [00:33:26]: 

Yeah. So the flip side of that is, of course, people are like, well, what am I gonna do all day? Like, what, what is my work gonna do? So as you go into organizations that I suspect are both excited about what AI brings, But also there’s some anxiety in that, right? There’s someone in that, or there’s someone in that organization that’s like, they’re not gonna need me anymore. How do you answer that question for them? What, what, how do you help them sort of see their place in the future with AI as part of the mix? 

 

Rachel Woods [00:34:00]: 

So another reason that I really love these hats that we’re outlining is, and I’ll say that the everyone hat is kind of the like, Just, you need to have everyone, you know, on board. But the other hats, they all really help people lean into their unique strengths and abilities inside the organization. And so I’ll give you a specific example of like the subject matter expert is probably the most common person to be like, well, wait a second. If I let AI now run my onboarding process that I was previously like the star of the show, being the only person that could run these audits or whatever. Then what am I gonna do, right? Am I gonna be, you know, replaceable? What you wanna do is you wanna help that individual see that their job is really about figuring out what’s next. 

 

Drew McLellan [00:34:47]: 

Hmm. 

 

Rachel Woods [00:34:48]: 

Their job is now about the world is your oyster. Maybe you think a little bit more like you’re like an R&D person. You’re inventing, you’re, you’re seeing how things are going and you’re figuring out what should we be doing next? Because there is always something else that can be built, done, improved that’s going to drive the business forward. And if you take somebody who especially is an expert at what they do and you can get them really like unleashed on the possibilities and their, their time is freed up because you have AI doing what they had been doing, 

 

Drew McLellan [00:35:22]: 

right? Yeah. 

 

Rachel Woods [00:35:22]: 

That person will unlock completely new things that are gonna, again, also help You get the result you’re looking for. 

 

Drew McLellan [00:35:28]: 

And connect dots in a way that only humans can connect dots. And if you’re, if you’re unburdened from the mundane work and you have time to stop back and look, that’s, I think, the other place that the human genius comes in. And all of a sudden, to your point, not only do you see the next step or what we could do, but you’re like, oh, there’s an interesting pattern that I’ve never seen before. wonder what we can do with that over here, right? 

 

Rachel Woods [00:35:56]: 

Exactly. That’s also why, again, mission critical that everybody on the team knows how to do this at a basic level. 

 

Drew McLellan [00:36:02]: 

Right. 

 

Rachel Woods [00:36:03]: 

Because what this tactically needs to look like is that subject matter expert going and, you know, having some time for you on a Friday and be like, well, I’ll give you a recent example for our team. Our, one of our, uh, sales reps was like, I was thinking that we could actually take like this part of what we know about our leads and we could change our follow-up sequence with AI to, you know, match some of the questions that they have in this I already made a basic playbook of it. I already have a working prototype cuz I was, you know, I know how to do this and I was messing around. Could we talk about like maybe testing this in a bigger scale? And so now I have, again, a subject matter expert who is innovating and pushing the business forward, solving problems. But then now we can take that playbook that that person created and turn it into something that’s more operational, right? Or, you know, operator implementer. Installing it inside how we run our whole system and make it the real thing. But that wouldn’t have happened without freeing up that person’s time. 

 

Drew McLellan [00:37:00]: 

Yeah. And brainpower. Right. 

 

Rachel Woods [00:37:02]: 

And, uh, you know, helping them see that that’s now their opportunity. Right. 

 

Drew McLellan [00:37:07]: 

Yeah. Rather than being afraid of it, they see, oh, I can, I can make it even better. So when an SME comes up with an idea like that, does it then, and, and they’ve built a little prototype or whatever. Does it then get shot back up to the operator to sort of operationalize it for the whole organization? 

 

Rachel Woods [00:37:23]: 

That’s what we recommend. 

 

Drew McLellan [00:37:25]: 

Okay. 

 

Rachel Woods [00:37:25]: 

Yeah. 

 

Drew McLellan [00:37:26]: 

Okay. All right. So one of the things that was interesting in Paul, we did a workshop with Paul yesterday and he was asking people, how many people primarily use this tool and that tool? And, you know, like he said, ChatGPT, like 4 people raised their hand. He said Claude, like everybody, you know, both arms are up in the air. So one of the things that, that you say a lot is own the playbook and rent the tech. Does it matter which of these tools an agency sort of locks into? And typically when you go in and help an agency, are they mostly in one space, whatever, whichever one it may be depending on their ecosystem, or do you have them doing things in all kinds of different tools? 

 

Rachel Woods [00:38:07]: 

So I love this question. Because actually, if most of this conversation has made you feel a little bit, you know, potentially overwhelmed of all of these new things now, let me help you take things off your plate, which is that you should really, as a business, only need one AI tool, and it doesn’t really matter which one. I like to, uh, look at all these AI tools, and if you zoom out, you know, 3 to 6 months, they all have all of the same features all the time. And it’s because they’re, they’re competing. They’re like, uh, they’re like jealous sisters, right? Like, well, if you’re gonna get an A in class, I’m gonna get an A in class, you know? And so it really doesn’t matter. And so if you are at the place where you have a bunch of tools, one of the highest leverage things you can do is just simplify down to one tool because now your team has the same shared language and people can start, uh, feeling like this is a team sport instead of like, oh, you’re using Grokbot, should I be switching over and wasting time learning Grokbot? Or maybe I just, they just mentally, you know, uh, thinking about switching over, right? So you’re wa— you’re wasting time with, uh, being spread out across tools. If you have to pick and you’re kind of torn between like which tools to use, my hot take is that you should go choose the tool that is most comfortable and common By the most people inside your company, not the most advanced. 

 

Drew McLellan [00:39:35]: 

Hmm. Okay. 

 

Rachel Woods [00:39:37]: 

And I know that’s gonna make you feel like, you know, especially if you’re using— 

 

Drew McLellan [00:39:39]: 

To your point, they’re all advancing. Well, first of all, every freaking day they’re launching a new version. 

 

Rachel Woods [00:39:45]: 

Google Gemini just said now they’re rolling out skills, right? 

 

Drew McLellan [00:39:48]: 

And so what I, what I’m hearing you say is they’re all kind of the same. Yeah. And so pick one. Don’t worry if it’s a little behind or a little ahead, cuz It’ll either catch up or the others will catch up to it. It’s really about just the consistency and comfort of the tool, you know? 

 

Rachel Woods [00:40:06]: 

And then being able to take your time and spend it on what really matters, which is the playbooks. And so, um, I know we’ve been throwing around the term playbook a little bit, but just to give a definition, really it’s a process that instead of a human having to do that process, it’s an AI doing that process. Um, and this is like a, a, you know, a methodology term that we use, but I’m gonna keep saying playbooks. Right. Yeah. So just figure. Uh, give the definition because then if you’re spending your time instead of testing out the new feature and blah, blah, blah tool, if you’re instead spending your time figuring out, okay, what’s the new process or, uh, playbook that we want AI to follow? In the worst case scenario, if you like, like, you know, uh, Armageddon breaks loose and you have to switch tools. Any AI tool that can follow instructions— 

 

Drew McLellan [00:40:57]: 

Can write the instructions, right? 

 

Rachel Woods [00:40:59]: 

Can, can write them, but also any, any AI tool that can follow instructions can follow a process that you’ve documented. So let’s say that you have all your playbooks written out in custom GPTs and they are going away now, right? Great. You just move that, that same stuff into skills. 

 

Drew McLellan [00:41:15]: 

And, and again, the way you do that is you ask the custom GPT to document itself. In a way that you could give it to another tool to do? How, how would you, how would, let’s, so the, they’re going away and a lot of people have them. So what’s the 1, 2, 3 way of going, oh, I need to move that into Claude or Perplexity or Gemini or whatever the tool is? 

 

Rachel Woods [00:41:36]: 

So it’s a little bit like you’re having, you know, like the employee on your team that you kind of just like trained like haphazardly and like randomly. 

 

Drew McLellan [00:41:46]: 

Oh, what agencies have done that? 

 

Rachel Woods [00:41:47]: 

Have you ever tried to get that person to then document their own work? Oftentimes it’s kind of hard, 

 

Drew McLellan [00:41:54]: 

right? 

 

Rachel Woods [00:41:55]: 

And what’s better is to start from the fundamentals of like, hey, let’s like look at this job description. Let’s figure out what are the different processes and let’s, again, this used to be really slow, which is why people didn’t do it. Now it’s fast, so there’s really no reason not to do it. But starting from the fundamentals of like, what are the different processes that we are having this GPT do? And then you should actually have the AI and yourself go through and document step by step what that process is. 

 

Danielle McCier [00:42:20]: 

Right. 

 

Rachel Woods [00:42:21]: 

And again, it depends on the types, cuz there’s also a million different ways you can use GPTs, but I would maybe look at it more like a warning sign that if you feel like you don’t have a clear path of how to migrate your GPT over into a skill, then you were not owning the playbook, renting the tech. You were just building into the tech. And the same thing is probably happening right now. If your main way of using AI is that you get it to remember in its memory or write a bunch of random skills that you’re not really sure how they work, you’re not owning the playbook and renting the tech. You are locking yourself into a tech. Does that make sense? 

 

Drew McLellan [00:42:58]: 

Yeah, totally. So as you were talking, I was thinking, so you’re building out all of these playbooks. How do you recommend for your clients that they keep track of everything they’ve built so that when a new employee comes on or something happens, they have this inventory and they know what are, here are all the tools in the toolbox. 

 

Rachel Woods [00:43:20]: 

That’s why this is an ops job. 

 

Drew McLellan [00:43:22]: 

Right. 

 

Rachel Woods [00:43:22]: 

Yeah, it’s the same thing with, uh, um, an SOP library, right? And so tactically what we do, and just to Uh, kind of boil it down. So we actually write playbooks outside of the tool. So we write our playbooks inside Google Documents. 

 

Drew McLellan [00:43:38]: 

Really? 

 

Rachel Woods [00:43:39]: 

Yeah. 

 

Drew McLellan [00:43:40]: 

Okay. 

 

Rachel Woods [00:43:41]: 

So then you have, or you can do it in Notion, or you can do it in, you know, your Microsoft. It doesn’t really matter. Some of our team internally, we’re also testing out and running our, writing our playbooks in GitHub. If anybody’s familiar with that, it’s just a little bit more of a technical path. Google Docs is Perfectly fine. And the reason you can do that is because then once you’ve written them out in Google Docs, and yes, you can have AI help you write these, but you want to really have it set and established because then you can just run it inside the AI tool, whether it’s like, you know, file download, now it’s saved as skill and that’s what’s running, or you can have it read from the Google Doc and run it. But if you have it separate, now you have the opportunity to have way more organization around it. 

 

Drew McLellan [00:44:25]: 

Well, and you’re tool agnostic, right? 

 

Rachel Woods [00:44:27]: 

Yeah. 

 

Drew McLellan [00:44:27]: 

Right? 

 

Rachel Woods [00:44:28]: 

Yeah. And so, um, when you think about how to keep track of stuff, can really be as simple as, we’re talking one Google Drive folder that has your different playbooks that your company has built. And it has all the version control also of Google Docs, which is huge, especially as you’re trying to have this be something that you’re operating with. Yeah. You’re gonna, you know, Jimmy’s gonna wanna change the playbook because Jimmy has a new idea. Great, let’s change it. But then if it doesn’t really work, we always have the way to go back. If you’re just building it inside the tool, inside skills, you don’t have that. 

 

Drew McLellan [00:45:04]: 

Yeah. 

 

Rachel Woods [00:45:04]: 

Yeah. 

 

Drew McLellan [00:45:04]: 

Huh. Low-tech, low-tech solution, right? I like it. I like it. All right. So if an agency does this really well and they’ve got in Google Docs or whatever, they’ve got this playbook library, let’s say, right? Does that become IP for them and is that an asset? So we do a lot of M&A work. We help agencies buy, be bought, or buy other agencies as an acquisition strategy or growth strategy, or get, find a buyer and sell when the founder is ready to get out, right? You’ve, you’ve walked through that before. Does this become part of their IP set and does it add value, do you think, to the business? What do you think? Yeah, I think so. I mean, uh, you know, we do a lot of work with, uh, all our listeners will be familiar with Sharon Torek, the IP attorney. And she talks about how a lot of agencies have all this IP that they’re not even aware that they have. But this to me would be like, I have this whole library of how to run a business faster, more efficiently, more effectively. Yeah, I would think that that would be a beautiful piece of IP that you could either sell separate. Or you could sell with your business, right? Totally. And so are you seeing agencies think about how to monetize that or how to, how to assign a value to that? Or are we too early yet to do that? 

 

Rachel Woods [00:46:22]: 

So, you know, in the M&A world, which I’ll say you definitely have a much, probably a more intimate seat than I do, I have at least been hearing the conversation and the way that it’s showing up is that Uh, having good SOPs and process or like a well-operated agency, that was always something that was going to improve the value. You know, it’s a value adder, not a value detractor. For sure. 

 

Drew McLellan [00:46:47]: 

For sure. It’s, it’s on our list of things like before you go to sell, you gotta get your ducks in a row for sure. 

 

Rachel Woods [00:46:53]: 

Yeah. It’s also probably impacting your EBITDA, right? It’s, it’s, it’s already across the, um, scorecard, if you will, of what makes a good and valuable business. 

 

Drew McLellan [00:47:04]: 

Yep. 

 

Rachel Woods [00:47:04]: 

Whether it by itself is something that people are specifically buying, I think it’s a really niche buyer probably at this point in time. But I would imagine we’re gonna be in a place where, why the heck would you buy a business that doesn’t have this? 

 

Drew McLellan [00:47:18]: 

Right. 

 

Rachel Woods [00:47:18]: 

When there’s a business that does. 

 

Drew McLellan [00:47:19]: 

It becomes a detriment. 

 

Rachel Woods [00:47:20]: 

Yes. 

 

Drew McLellan [00:47:21]: 

Yeah. 

 

Rachel Woods [00:47:21]: 

Because it’s not just a, a well-operated SOP library anymore. It’s a way to infinitely scale. 

 

Drew McLellan [00:47:30]: 

Right. 

 

Rachel Woods [00:47:30]: 

The labor inside the company in a trusted, reliable way because you have AI following all of these playbooks and processes. 

 

Drew McLellan [00:47:39]: 

Yeah. 

 

Rachel Woods [00:47:40]: 

And so if tomorrow I wanted to buy a TikTok shop agency and they already have everything playbooked out, let’s say that that’s a tiny TikTok shop agency and I’m a behemoth and my plan was to integrate that talent so that I can now offer TikTok shop to somebody else. Well, in the old way, the immediate problem is like, okay, I’m gonna have to hire more people. Right. And train them on their SOPs. You don’t have that anymore. You can immediately unlock the value of these playbooks in a much bigger scale if you have that demand. So those are the types of things to think about. I think especially if you are a smaller, uh, business, I mean on both sides, but where you can have really outsized value to an acquirer because 

 

Drew McLellan [00:48:22]: 

Yeah, it’s exponential. 

 

Rachel Woods [00:48:23]: 

Exactly. 

 

Drew McLellan [00:48:23]: 

Yeah. Yeah. Okay. So we’re talking about money. So let’s, let’s get to the scary part of AI for a lot of agency owners, which is, and you know, from the beginning of time, agencies, you know, back in the Mad Men days, we, you know, everything was based on the billable hour or the media commission. Now fewer and fewer agencies are buying media. So then it went to billable hours, like, oh, you need a thing, it’s gonna be 12 hours. Each hour is X number of dollars. Now agencies have kind of evolved to, oh, you need the thing. Here’s a flat fee project. But underneath that flat fee project are, I’ve done the calculation of how many hours it’s gonna take. So I think one of the big fears for agency owners is as, as the world and our world implements AI and our clients get savvier to what AI makes possible. Mm-hmm. How is that going to impact my ability to bill a client? So you and I were talking before we hit record and it was like, you know, I thought of a great tagline in the shower. Do I charge the client just the 15 minutes for how long it took me to take the shower or do I attach another fee? So how are you helping agencies think about kind of the P&L of this and how they answer the question that a client is going to say, well, Did you use AI on this and how much time did it save you? And therefore, why am I paying that much for the thing? 

 

Rachel Woods [00:49:49]: 

So there’s actually a really simple framework that I think you have to kind of pick which boat you’re gonna be in, right? And it comes down to really where does the arbitrage of using AI live, 

 

Drew McLellan [00:50:03]: 

right? 

 

Rachel Woods [00:50:04]: 

And ultimately, where does the responsibility of the outcome live associated to that? So let me break this down. If you have a business, uh, an, you know, an agency where you’re building out, let’s use, you’re doing like founder brand content for your clients. I, I’ll use this cuz we actually just hired an agency that’s, uh, doing some stuff with us on this. So, uh, you have in the AI world, I think one of two options of how you structure your engagements if you’re going to use AI in your work, which you absolutely should. One is that you are passing on the arbitrage, the time savings, the money savings to the end customer. How do you do that? Well, you need to decrease your costs, and it means you also need to probably give them direct access to the AI, right? A lot of agencies are like, hey, we, we set up like a bot that now you can go and build out content and use— 

 

Drew McLellan [00:51:00]: 

Blog posts or whatever. 

 

Rachel Woods [00:51:01]: 

Yeah. And use our, um, Our special sauce, right? If that’s the model that you choose, again, you need, you need to pass on the cost savings to them cuz they’re now the ones with the responsibility of the outcome of like how good the work is because they’re the ones using it. 

 

Drew McLellan [00:51:14]: 

Right. Yeah. Right. They’re, they’re the human overlay. 

 

Rachel Woods [00:51:16]: 

Yes, exactly. 

 

Drew McLellan [00:51:17]: 

Right. 

 

Rachel Woods [00:51:17]: 

But the other option is to, again, be upfront that you’re using AI, but you are still holding the responsibility of the outcome, right? And the quality of work, meaning that your team’s using AI, but That, that client doesn’t see it, doesn’t care, doesn’t know how much and where and how much you had to edit. Like they don’t, that’s none of their, uh, in their purview because at the end of the day, it’s your team getting on the call with that client, delivering an outcome of a certain quality, and your team is responsible for that. That world is, I think, the smarter world to be in if you have a talented team. 

 

Drew McLellan [00:51:52]: 

Right. 

 

Rachel Woods [00:51:52]: 

If not, then you’re competing in the, SaaS environment, really. Right. And so, in that world, again, I think you should always be trying to charge based on the value you’re providing. The value should be the same, if not more. So, if you’re really concerned about this, an easy thing is to figure out how can you deliver even more for your clients with AI? And I guarantee you, they’re not gonna complain that you’re delivering more, and that’s gonna be the conversation, not like, should the prices go down, right? But the important thing is to make sure that you’re still holding the, you know, the responsibility of the outcome, and that’s what earns you the right to have than the arbitrage of the time. 

 

Drew McLellan [00:52:27]: 

So you and I are sitting across the table, you’re the agency president, head of sales, whatever. And I say to you, wow, $35,000. Aren’t you guys using AI? Couldn’t you do this faster and therefore less expensive? What is, what’s the script? So I completely agree with what you just said, but what’s the, what’s the client-facing script? 

 

Rachel Woods [00:52:53]: 

So I think it’s, it’s the same script that you used pre-AI on any pricing objection, which is what? Go back to the value. 

 

Drew McLellan [00:52:59]: 

Right. 

 

Rachel Woods [00:53:00]: 

Like you should never be answering a pricing objection with how much time or the features of what you offer. It should be about the value that you’re providing and the cost of inaction for the customer. I mean, this is just a sales thing. This isn’t, isn’t as existential, I think, as people think. This is just a sales thing. 

 

Drew McLellan [00:53:19]: 

Sales comfort thing, right? 

 

Rachel Woods [00:53:20]: 

And it’s, um, I love, I love sales objections that are actually just one of the common core objections, right? Of, of money and trust, but they parade as something else. This is just the money and trust objection with an AI flavor. So handle it the exact same way. 

 

Drew McLellan [00:53:35]: 

So I believe, and I’m curious if you believe, uh, so, uh, Danielle and I do a big trends report every year, and one of the things that we were saying is I believe this is the season for agencies to be more profitable than ever they’ve ever been able to be before, because they are going to be able to do more in every hour of human work with the support of AI. The output and the value of the output is going to be even better. So in theory, a smart agency who uses AI wisely the way we’ve been talking about actually should be more profitable than ever before. And clients still happy cuz they’re getting as good, if not better outputs, right? 

 

Rachel Woods [00:54:18]: 

I agree. And where I’m seeing this is, um, you know, the, the accidental agency founder? 

 

Drew McLellan [00:54:25]: 

No, all of ’em accidental business owners. 

 

Rachel Woods [00:54:28]: 

That’s, uh, usually like, you know, incredibly good at your craft. And what, what used to be like the old, the old way had such a, an immediate bottleneck that you would hit, which is that you can’t scale your own time. So now you can’t just do your craft. You gotta hire and manage and figure out all this stuff. And that’s why agencies tend to be small, right? Until they figure out how to break through that. You don’t have that problem anymore. So some of my favorite agencies are literally people I actually just have a, an owner I’m really close to that broke up with their business owner. I definitely, uh, can’t say the details of this, but I think this is a crazy story. Broke up with their, uh, previous co-founder because they’ve been, uh, not really adopting AI as fully as they could. And there were some disagreements on how aggressive they should be about it. But this guy is a, um, very talented, right? And so what’d he do? He went out with 2 other people in his network, also very talented, all AI savvy. And now they are scaling. I think, uh, last I talked to him, they’re about 5 or 6 months into the journey. And they’re already doing more revenue per month than their old agency had been doing that had been around for, with all the mergers and stuff, like 6 or 7 years. They’re 5 or 6 months in doing more revenue with a 3-person team because they have AI systems scaling themselves and they’re good at what they do. And so if you’re an agency owner running a more traditional agency, I think that’s actually more the risk to be concerned about. 

 

Drew McLellan [00:56:00]: 

Right. 

 

Rachel Woods [00:56:00]: 

Than the pricing conversation about we’re using AI or not. It’s, are you gonna be disrupted by like brilliant people that now can be completely unlocked and scaling without having the, uh, previous kind of like the old way constraints? And so then we get back to the hats, 

 

Drew McLellan [00:56:20]: 

right? 

 

Rachel Woods [00:56:20]: 

That’s your subject matter experts that you wanna make sure feel incredibly bought in and incredibly empowered to keep building and iterating and innovating inside your business. Because if you don’t have that, good chance they’re gonna go do that. 

 

Drew McLellan [00:56:36]: 

Right. It, it, it is. We, we’ve been saying, you know, that we as agency owners need to be figuring out what is the Uber that’s gonna put our taxicab company outta business. We don’t, and, and we can be the Uber, right? And that’s different than I think other industries, but we can be the disruptor. Rather than be disrupted. If we are willing to sort of embrace all of this and figure out how to better serve our clients, which ultimately is still, it’s still about the craft of the creativity and the marketing strategy and the all, like that doesn’t go away. We just have new tools. 

 

Rachel Woods [00:57:19]: 

And you know, when I was, um, building my agency and I was thinking about You know, the, the common advice, right, is like, pick the customers you wanna serve wisely and really think about like who you love spending time with, who you love helping, who can have outsized results. Agencies fit all of those things. And the number one thing that you need to remember is that I think a lot of what an agency is, is it is the innovation hub of business, 

 

Drew McLellan [00:57:46]: 

right? Yeah. We’re always the, we’re always out ahead. 

 

Rachel Woods [00:57:49]: 

Yeah. 

 

Drew McLellan [00:57:50]: 

Right. 

 

Rachel Woods [00:57:50]: 

And so you may feel, again, back to, you may feel out of like out ahead right now because you kind of check the box of like you’re using AI and you have a couple use cases. But hopefully this conversation has given you some ideas of how, where you should be next, right? Because the, I like to say, one of my favorite quotes is, um, that, you know, the future is here. It’s just not evenly distributed yet. And there are organizations in your industry or in adjacent industries absolutely running this way. And there’s more and more pop up every day. So that’s the opportunity is to still be ahead and keep pushing. Don’t settle. 

 

Drew McLellan [00:58:27]: 

Yeah. You know, we work with hundreds of agencies every year and I can tell you the range runs from owners afraid and maybe has a ChatGPT account and people underneath him or her are doing things, but it’s very individualized because of the fear at the top. And then we do have agencies that are. Following sort of your recommendations and your methodology and, and they have changed the way operationally they run the business. So you’re right, there’s a huge spectrum. So you got taxicab on one side and you got Uber on the other side. And, and so just having an account isn’t really the forward thinking that maybe they feel like it is. 

 

Rachel Woods [00:59:08]: 

Yeah. And then the other piece, just to introduce Kind of to the reality is I think, especially in the agency world, talent acquisition and talent retention is everything. Yeah. And a huge chunk of our business is that we train individuals that just come straight to us. And the number one reason is they think that their employer’s not taking this seriously and they don’t wanna be left behind and they don’t want their skillset and the pace of their progress to be defined by the pace of their employer. 

 

Drew McLellan [00:59:39]: 

No, that’s, yeah, right. That’s scary. 

 

Rachel Woods [00:59:40]: 

Yeah. 

 

Drew McLellan [00:59:41]: 

Right. 

 

Rachel Woods [00:59:42]: 

‘Cause imagine if you’re somebody and, and you’re looking and you’re like, my company’s just not taking this seriously and I don’t really feel like this is a thing I can just learn overnight if I were to get laid off or like something were to happen. And those people, they’re also looking out on the job market of, um, there’s data that PwC has published that jobs requiring AI skills like we’re talking about are growing 8 times faster than the overall job market. 

 

Drew McLellan [01:00:05]: 

Right. 

 

Rachel Woods [01:00:05]: 

And on average in 2026, they’re paying 62% more on average. So a $100K job is paying $162,000 a year on average, just the same job, the same seniority, the difference being AI skillset. And so if you are inside a team that’s not taking this seriously or not investing in your development, then like that’s gonna keep you up at night. 

 

Drew McLellan [01:00:29]: 

Well, and to your point, those people are going to leave and find new jobs. Which means you’re gonna be left with the B and C players. 

 

Rachel Woods [01:00:37]: 

Exactly. Yeah. 

 

Drew McLellan [01:00:38]: 

So again, even further behind. 

 

Rachel Woods [01:00:41]: 

And, and entirely, I love, I love knowing problems when they’re entirely preventable. That’s like the best kind, right? Right. This is one of those entirely preventable. 

 

Drew McLellan [01:00:49]: 

Yeah. All right. So if somebody’s listening, as we, as we kind of wrap up this conversation, you and I know that I’ve gone through about a third of the questions I had to ask you, but as we wrap up this conversation, somebody’s listening and going, Crap, I gotta get on this. I, I really, I thought I was ahead, but I really am realizing now I’m not. I’m either wearing all the hats or I’m, I’ve got a hat and a couple of my people have. What do they, what’s first step? What is first step on Monday morning and how do I start this evolution? 

 

Rachel Woods [01:01:22]: 

So the easiest first win is really To, this is going to sound crazy, but you need to temporarily put project ROI out of your mind as how you evaluate AI ideas, AI initiatives. You need to temporarily put that out of your mind. You need to focus on getting everybody to see this as a new way to work. And you need to have an appetite for, I mean, depending on how much time you want to spend, we usually say it takes about 15 hours to take somebody from using ChatGPT to being able to playbook their work Pretty well. 

 

Drew McLellan [01:01:58]: 

Wow. 

 

Rachel Woods [01:01:58]: 

15 hours of work. 

 

Drew McLellan [01:01:59]: 

But that’s not a lot. 

 

Rachel Woods [01:02:00]: 

It’s not a lot. 

 

Drew McLellan [01:02:01]: 

Right. 

 

Rachel Woods [01:02:01]: 

But in an agency, I mean, sometimes that, that can be hard to find, right? Because you have, uh, packed calendars. But that’s one of our programs is we run a launch program that, uh, it takes 15 hours for people to get from there to there. But the best next thing you can do is get everybody on your team to be able to take something in their work and structure it as a process that they could hand off to AI. Where they actually now feel like they’re building up kind of the operating rhythm. You need momentum because that momentum, everybody now feeling like, oh, okay, I kind of get this. I, I get where we’re going. I’m bought in. That then helps you now pick the high ROI projects and focus the energy and effort to then work on the harder stuff. But right now, if every new AI idea is coming to your table and your first question is like, what’s the ROI on this? You are shutting your team down. You need to instead get them, yeah, bought in in the momentum. 

 

Drew McLellan [01:02:52]: 

Get in the sandbox and play. 

 

Rachel Woods [01:02:53]: 

Yeah. 

 

Drew McLellan [01:02:53]: 

Yeah. 

 

Rachel Woods [01:02:54]: 

In a structured way. Yeah. Where there’s outcomes. Right. And, you know, it’s clear, and I’m not, I’m not saying like, just take 15 hours to go do whatever, but a clear outcome is like, everybody should have at least one playbook in their work, a process that now AI is doing. Um, and that is a very easy way to start getting this embedded into your operations. 

 

Drew McLellan [01:03:13]: 

Awesome. I know a lot of people listening are gonna wanna learn more about your company. They’re gonna wanna follow you. They’re gonna wanna learn more about the work you do. So what’s the best way for them to do that? 

 

Rachel Woods [01:03:23]: 

So 2 things. One, I’m super active on LinkedIn, so message me there. And we do run a free class on how to do this playbook approach. I know we didn’t talk in depth about, you know, the triggers, the inputs, the steps, the outputs, how to write these things, where to save ’em, all that stuff. Teach a free class that I’d recommend, you know, you and a couple people on your leadership team, or maybe even People that you’re like, these might be our AI operators, right? Leading more of this. Um, come join one of those. We do ’em, uh, at least once a month, and that’d be a great place to start. 

 

Drew McLellan [01:03:52]: 

So where do they go to find all of this other than following you on LinkedIn? Where else do they go? Where do they, where do you want them to go? 

 

Rachel Woods [01:03:58]: 

Uh, you can message me on LinkedIn or you can go to runonamp.com. 

 

Drew McLellan [01:04:02]: 

Okay. We’ll make sure that’s in the show notes. This has been an awesome conversation. I am so glad we got to do this. Thanks so much for being on the show. 

 

Rachel Woods [01:04:09]: 

It’s been really fun. You asked such good questions. It’s been awesome. Yeah. 

 

Drew McLellan [01:04:13]: 

All right, guys. So this, you know, I like episodes that are action-packed, that give you homework, and boy, did Rachel give you homework. And the first one I think is mindset. You have to figure out as an agency owner how you feel about AI, where AI plays into your agency, what you can do with AI to benefit your clients. Your team, your agency, and ultimately you and your family. Like the opportunities here are huge, but you gotta be in the right mind frame to do that. So that’s homework 1. And then homework 2 is go take this class. I mean, that seems pretty obvious and easiest. And then do what Rachel said, take the 15 hours, have everybody build something out. But this is not optional. I, I’ve told you all the story of Worked with a guy back when computers were coming into the agency, which I know ages me. He was like, I’ll never— he was an art director. I’ll never do anything on a computer, blah, blah, blah. I’m gonna do it all by hand. And we had his retirement party 6 months later. And that’s, that’s the wall that we’re running up against. The good news is you have everything you need to scale that wall and to grow your agency in a way that you can’t even imagine yet to the benefit of everybody in your world. And so. Homework for you is get, get your head on straight, surround yourself with good resources. We have some, you’ve got Rachel now, you’ve got lots of other agency owners that, you know, are doing things that maybe you haven’t done yet, but surround yourself with people who see this as the opportunity that it is. And then get to work because, because it’s time for us to rebuild our industry with these tools, which is pretty exciting. Okay. All right. Thank you for listening. I am always super grateful that you hang out with me every week. I’ll be back next week with another great guest. And until then, talk to you soon. 

 

Danyel McLellan [01:06:04]: 

That’s a wrap for this week’s episode of Build a Better Agency. Visit agencymanagementinstitute.com to check out our workshops, coaching and consulting packages, and all the other ways we serve agencies just like yours. Thanks for listening.