333 | Stop Prompting. Start Playbooking: How to Automate Your Business with AI with Rachel Woods

0:00Isar Meitis Hello, and welcome to the Leveraging AI Podcast, the podcast that shares practical, ethical ways to leverage AI to improve efficiency, grow your business, and advance your career. And I have a really awesome, amazing, incredibly useful episode for you today. Everyone, literally everyone who's using AI, which is everyone, wants to automate things in their business with AI.

0:21Isar Meitis Now, some people actually have one-off solutions of different aspects of automations that they were able to build, but the vast majority of people don't even know where to start when it comes to automating things with AI. Most people use it as a glorified, Google search or maybe a little more, but never really dove into how to actually make AI do things in your business.

0:42Isar Meitis Now, the biggest gap I believe that people have between, "I know what the problem that I'm trying to solve," to, "How do I actually implement AI so it actually solves that problem for me?" is that people don't know how to accurately and effectively define the actual requirements. What is it exactly that the AI needs to do in order to help me do this thing that I'm trying to do in my business in a more effective way?

1:05Isar Meitis Now, since AI is incredibly good at implementing whatever you tell it to implement, defining the requirements is the biggest issue. If you know how to effectively define the requirements, then you're already 80% on your way to success because the AI will know how to implement these requirements, probably by asking you some follow-up questions.

1:28Isar Meitis Now, this is exactly what we're going to teach you today, how to create playbooks that are basically a, like the name suggests, it's a recipe on exactly what the AI needs to do in order to solve a specific part of the problem, and then you can take what you're gonna learn today and implement this in your business.

1:45Isar Meitis Now, our guest today, Rachel Woods, is first of all, an AI ninja. She's been very, very active in the space, more or less since I was, right? So the end of 2022 when ChatGPT came out. she's a serial entrepreneur. She started multiple businesses. Before starting all these businesses, she spent time at Facebook AI, so she also has the back end, more detailed technical knowledge and experience in how to implement AI, and that's pre, ChatGPT.

2:10Isar Meitis So she's been in that space for a very long time. But she's been extremely active in teaching people and individuals on how to implement AI in their businesses in an effective way. She has three different companies that are all involved in doing this in different, ways, and she's going to show us step by step exactly what is a playbook and how you can create your own, and how you can implement one in order to automate things in your business.

2:33Isar Meitis Now, Rachel and I have known each other for a while now because we're swimming in the same lakes and we started doing very similar things, a long time ago, again, roughly at the same time. So we're coming up on four years. That's crazy. So about four years, and we talked several times before, but we never actually had a chance to collaborate, and then we ran into each other speaking at a conference a couple of months ago, like, "Oh, we gotta do this," and like, "Yes!"

2:54Isar Meitis And so we're finally here, so I'm really humbled and excited, to have you on the show. I know that what you're gonna share, a lot of people will find extremely attractive, and so welcome to Leveraging AI, Rachel.

3:08Rachel Woods Thank you. I do feel like this is, quite overdue, isn't it?

3:11Isar Meitis It is. It is. But, but again, we're here. That's what matters, right?

3:14Rachel Woods We're here. We're here. The, the important things always find a way to, come about, right? So-

3:19Isar Meitis Yeah

3:20Rachel Woods I'm, I'm excited for this. This is gonna be really fun today.

3:22Isar Meitis Yeah, me too. I, again, you I think you and I think in a very similar way. I think we approach these things in a very structured, process-oriented way, and I think that's what most people miss. Most people are trying to wing it, and you can wing one or two, and it's gonna be very painful, but if you wanna automate your business, you can't wing it.

3:44Isar Meitis It has to be very much process driven and very structured, and in that, you and I approach it in a very similar way conceptually. probably very different ways on the tactical level, which will make this really fun.

3:56Rachel Woods Yeah. that's a nice way of saying that we're a little bit of, like, ops nerds. Aren't we? Yeah.

4:01Isar Meitis 100%, 100%.

4:03Rachel Woods But, I think, it's about time that operations is, is considered sexy, so there we go.

4:08Isar Meitis I'm with you. Revenge of the nerds.

4:11Rachel Woods Yeah. yeah, where should we jump in?

4:13Isar Meitis so let's start I, I know you've kind of like, y- your main thing is those playbooks, and I think I, I used it in the intro, and I've saw you speak about this many times. So let's talk about what the hell is a playbook, and why should people care so people understand why they should stick with us for the next 35, 40 minutes, and then we can dive into the actual tactical details.

4:32Rachel Woods Cool. That sounds good. Yeah. as you alluded to, you know, there's a, a ton of different ways to use AI, and, the last thing I want people to feel is that, like, um, the way you're using it is, quote-unquote, "wrong." there's just so many levels to it, and what I hope today is that we can introduce you to a new level that you might already, not already be using or thinking about, which is, really trying to think about AI not just as something that you're chatting with or using to get, answers or searching or researching, but really how to use it as an operational tool, right?

5:00Rachel Woods And, again, when I think about operations, it's really about being organized in your work so that then you get things done the same way or the way that you want every single time. And bringing that type of mindset to AI is actually, to me, is really the pathway to unlocking AI as something that you can trust and you can actually start to hand off a lot more work to it.

5:24Rachel Woods Not too dissimilar from normal operations, right? Of, like, building out a team and training people and, you s- you build operations so you can trust it. So at a high level, that's really what playbooking is it's a methodology that we have, but in the goal of creating, AI operations.

5:37Isar Meitis Awesome. Yeah, I think you touched on the key point, right? Consistency. Like, you want consistent results regardless if it's a human operation or a AI opera- whatever other operation. If you wanna run a business, the results have to be consistent, meaning they have to deliver the same kind of output, whether it is a product, a service, a something, and it has to deliver it the same way at the required quality- Right every single time, regardless on who's the user, who's the operator, whether it's AI or human or whatever.

6:05Isar Meitis And if you can build your environment to do that, then you can scale it.

6:10Rachel Woods And I think it's also, something that helps you move up the spectrum, right? Of 'Cause we all ideally want this AI employee, agent, whatever you wanna call it, in the background that's just doing all of this work for us, we never have to check on it, and we trust it's doing a good job. you know, but if anybody's tried to set some of those up and, been disappointed, where it's like, man, the longer the leash that I give it, the more chaos it creates, right?

6:32Rachel Woods I don't know if you've faced that, but, I certainly have whenever I get a little bit, a little bit too, lazy or fast or am- ambitious about, like, oh, I'll just have the agent go off and do all this stuff. It's like, man, you really, it really bites you in the butt pretty quickly, right? instead, when you give it a longer leash, if you can give it that, but also give it the structure so that it's, either doing what you told it to do or it's helping to create a standard way of doing something, that's what's gonna help you actually, again, hand off more work to it, and then it be able to end up with a deliverable that you actually want.

7:03Rachel Woods and I think that's, a lot of us, again, are kind of looking to reach that next level, and that's really, the shift that me and my team have made of not just, playbooking or creating these processes and structures for, small tasks, but actually then, stacking them all together into a system, and then you have the playbook checking other playbooks, and then them all working together, and that's then I feel like you have, this whole department or whole team of AIs working for you.

7:25Rachel Woods And this, kind of method we use is really that path to get there.

7:29Speaker This episode is brought to you by the Multi-Agent Orchestration Course. It's a course that I have been teaching personally since April of this year, and every single one of the cohorts that we have opened have sold out. And the people who have been through these courses made incredible transformations for themselves, switching jobs, starting their own companies and so on, and definitely for their businesses by being able to build really complex processes by following our methodology and our infrastructure settings that we teach during the course.

7:57Speaker The course is four sessions of two to two and a half hours each over four consecutive weeks, and through it, you can go from knowing a little bit about how to use AI and using it in ways that most people are using it today to learning how to build really complex, sophisticated systems that can automate more or less everything in your business.

8:13Speaker The current cohort that is open starts on November 2nd, and it's gonna go on November 2nd, November 8th, November 16th, and November 23rd, and we are almost sold out for that cohort as well. It is very likely that we will not open another cohort in December, meaning the next one after that is gonna be in January.

8:29Speaker So if you still wanna learn how to build really incredible automations and learn everything that I've learned by doing this every single day for the last few years, don't miss this opportunity to come and join us on November 2nd. There's a link in the show notes, And if you're going to use the promo code LEVERAGINGAI100, you're going to get $100 off of the price that is on the website.

8:50Speaker Like I said, there's a few seats open, but they are selling out fast. So if you're interested in doing this still this year, go and click on the link right now on your phones and complete the registration and the payment so you can guarantee your seat. I'm looking forward to seeing you there. And now back to the episode

9:07Isar Meitis Awesome. I will add two cents and then I think we'll do two things. One, you'll give a few examples of what kind of playbooks you have so people can make it more tangible. Yep. And then we'll actually dive into what it is and how to use one and how to build one. But the two cents I wanna add is really have to do with what you said, and I think that's another critical aspect that you and I think very much the same.

9:25Isar Meitis You wanna think on a very granular level when you build these things. And the reason you wanna think on a granular level, they become Legos that then can be built together into bigger things. And it-- you wanna do this for two different purposes. Purpose number one, AI just works better that way. So when AI has a very simple, specific, well-defined task, it will do it amazingly well.

9:44Isar Meitis When you give it a very complex, long, ambiguous task, it may or may not do it great. Now, the reality is the big, more complex task is built out of a lot of smaller tasks. So if you can build these things to know how to deal with a smaller task, you can put them together in a way that will be able to do the entire process, or like Rachel said, have an entire team doing something that will also check the work and verify and will transfer the results to wherever it needs to be transferred and so on.

10:09Isar Meitis And so thinking about your process not as the final outcome, but what are the things that have to happen in between, is very helpful because you'll be able to use these in more than one way. The best example that I always like to give people, when people see my slides when I speak on stage, they're like, "Oh my God, how did you create these?"

10:24Isar Meitis I'm like, "I-- first of all, I don't create anything anymore." but there's multiple skills, not one that creates these slides. And the best example I like to give is my brand guidelines. everything has my colors and my fonts and blah, blah, blah, and whatever. I could have created a branded presentation skill or agent or whatever you wanna call it.

10:42Isar Meitis I didn't. There's separate ones. Each one does something different, so I can reuse the one that creates the graphics in other things that require graphics, like my website or pamphlets that I create, and one that creates my color and branding on wherever I wanna use this, and one that create presentation.

10:58Isar Meitis I can create presentations for my clients because it's gonna use the same presentation creation skill with other branding capabilities and so on. So being able to mix and match is very helpful once you wanna scale it beyond the simple thing that it's doing. Now back to you. A few examples, and then we'll take it from there.

11:14Rachel Woods Yeah. I mean, I think even just building on what you're talking about, because, what you're really describing is the idea that you can have individual processes inside a business, individual things that get done, that then also add up to a bigger, outcome, right? Which again, my- one of my favorite things to do with AI is just, forget AI.

11:33Rachel Woods Just think about normal operations, normal ways of building business, and if you were to think about, let's say you were to, I'll give an example. Hire and train an assistant for yourself, right? A human assistant. To help them be successful, you probably wouldn't, day one, just be like, "I want you to, do my entire morning routine, and I want you to, Here's, my inbox and my calendar and my, all my messages and all this stuff.

11:54Rachel Woods go." you wouldn't just kinda, dump it all. If you were trying to set that person up for success, you would probably think through, "Okay, what's, the first thing that I should teach this person? First small process I could get off my plate. Okay, the first thing I wanna teach them is how to give me a daily briefing of my calendar.

12:10Rachel Woods Great. Let me teach them that. That's, day one or two on the job Then day three you're like, "Okay, now let me teach the next thing," right? Which is "Here's how I want my email inbox managed." And then you would, teach them that. And you would keep going until eventually you're like, great, now in the mornings my assistant runs an entire morning routine for me because my assistant knows how to do all these different things.

12:31Rachel Woods Oh, and by the way, if for whatever reason we get to, like, you know, 1:00 PM on a Tuesday and my whole day has just been chaotic and I need a new daily brief 'cause my calendar completely changed, great, the assistant knows how to do a calendar brief in itself. It doesn't have to be constrained into the morning routine.

12:49Rachel Woods And so, like, that's just, normal building a team, training people kind of mindset. And what I love is, like, that's just exactly how you can approach breaking up your work and designing these systems for AI to do. AI loves the exact same things that training a human does. And even just to give you examples, those are literally examples of playbooks I have that power my day, and then my morning routine, I can start the whole thing, with just one word, which is todaying, and that goes through and triggers all of these different playbooks, just like I hired that assistant.

13:19Rachel Woods And so you can do this across anything. You can do it in your, your personal productivity. You can do it in your sales process, in the way you deliver your service. I'll give you another example, again, back to the, breaking apart work. if you think about any, service that you might have in your business or, let's even take something like customer onboarding, right?

13:35Rachel Woods If you were to systematize that, what do you do? You probably chunk it up into the different smaller pieces that are required for customer onboarding. If you playbook each of those things, then you have a playbook running the whole process together that then becomes now a very easy way. Again, same thing to get it to do bigger things.

13:53Rachel Woods I just really like thinking, again, about, if you were to teach a team of humans to do this, how would you break apart the work ideally? And then that's a great, jumping off point for AI as well.

14:02Isar Meitis Awesome. So through this, you gave us two small examples, like a morning routine. What other kind of playbooks? what would you consider a playbook, at the simple level, and then what would you consider a playbook at the slightly more advanced level?

14:14Rachel Woods Yeah. So actually to clarify, so even in the morning routine, there's actually a bunch of sub-playbooks, right? So there's like email triage is a playbook that I have. and the reason that it itself is a playbook is because, and I don't know if anybody probably can relate, like if you just give AI, like, damn, if you just give a person, "Triage my email inbox," you don't give it a lot of like clarity.

14:34Rachel Woods you're never gonna trust that thing. It's gonna go crazy. It's gonna try to do all this different stuff. So instead, my email triage playbook actually has, like a decision-making guide of like what are the different types of emails that I expect? What do you do if you need to escalate an email to me?

14:48Rachel Woods And so this playbook itself, I think, right now it's 11 steps, and it's because it's actually going through, pretty, pretty detailed. but that's how I can trust it, right? Because it's actually really following a real process. So that's one playbook. and then again, I have other playbooks, inside my morning routine.

15:04Rachel Woods But then, and I like to think of this almost as like the manager playbook, right? Which would be like the morning routine playbook is the one that delegates the work to the other playbooks, kinda a project manager in a team of people would do, or a chief of staff with-- if you had a bunch of different assistants, right?

15:20Rachel Woods and so that's an ex- another type of example of a playbook. But That's just in, yeah, personal productivity. the way that we run our business is everything is playbooked. So we actually have a system of playbooks for marketing, system of playbooks for sales, system of playbooks for, the delivery of our services, system of playbooks for even the leadership side, so how we manage strategy and reporting and everything.

15:43Rachel Woods And inside each of those systems is probably on average, I'd say, probably a dozen to three dozen different playbooks. and that really is, like, how we run everything in the business. We're 10 people, 10 humans, and, I think the last time I counted we were well over 100 playbooks powering in the back end.

15:58Rachel Woods So- Yeah I'm curious. W- out of those, what do you think... I can give some more examples, but what of those systems do you think would be, like, most interesting? Or what are you curious about?

16:04Isar Meitis No, I think this is very helpful. I think if we... Y- you already gave a lot of examples for people to understand what it is. I think we can now dive into the how. Like, I think now people are curious enough. They're like, "Okay, so I understand this whole playbook thing, but how do I actually... Like, show me one.

16:20Isar Meitis How do I actually create a playbook? What's in it?" And the other thing that I know the answer, I'm gonna ask you for the audience, is the creation of a playbook by itself, meaning is there a structured process on how to create a playbook that is the same every single time but creates different outcomes?

16:34Rachel Woods Yep. Yeah, great questions. And, of course you know the answer. Yes, of course there's a, a playbook for creating, playbooks. and so I really, I think what's helpful maybe is to bring back what you even just introduced at the beginning of this conversation of the hard part and the valuable part is in deciding what needs to be done, right?

16:53Rachel Woods The judgment, the structuring of the work. And that's really, the, the skill to hone, in my opinion, of people who want to use AI this way, is you need to build the skill of being able to break down any work. my team, makes me say this more often than I feel comfortable, but it is true, that I have not met a process I couldn't playbook because I have developed the skill of being able to break down literally any type of work.

17:16Rachel Woods And that really is the thing to lean into. So how do you do that? the way that we do it, again, this is like our method of creating these playbooks, is if you start by defining th- these same four components for any type of work, then you're going to have the hard part done, and then from there it becomes a lot easier to get AI to help you finish the rest.

17:34Rachel Woods But the first thing you need to figure out is what is gonna be the trigger Of this work. So what is gonna be the thing that starts the work? and that could be something like on a schedule, or it could be something like an event, like, when a meeting finishes, then start the playbook. But you need to figure out, like, when the heck are we even gonna, run this thing?

17:52Rachel Woods and that actually helps you really, orient around, the next step, which is that you need to figure out the inputs. So I think of, playbooks again, really, as, you're wanting AI to do something for you. sometimes it can just do that without you needing to give it anything, right? it might be able to go get what it needs itself.

18:09Rachel Woods But sometimes you actually need to give it something. So an example of that would be, like our newsletter writing playbook. that one, I actually give it a brain dump of what I wanna talk about because I want the topic and the substance to come from me, right? So that would be an example of an input. Once you have the trigger and the inputs, the third component you need to define is the steps.

18:28Rachel Woods So this is really, if you were to sit here and talk to me about, "Hey, here's how I want you to do this work, Rachel," then that is, exactly what you need to outline for, deciding, how you want AI to do something, right? Of just writing out those steps. And then the fourth piece is the output. So what do you expect to get out of the playbook and what's the deliverable?

18:47Rachel Woods And again, that helps be, like, a bookend, but, these things also help when you try to start stacking them together like we're talking about into bigger systems. But yeah, y- for any work, you just need to define those four things. Agreed and then once you do that, then you, depending on the type of work, you'll wanna, write out more in-depth instructions for some of these steps, right?

19:06Rachel Woods Sometimes it's really easy. a step might be, let's see, Get a full list of all of the tasks assigned to me in my task manager tool, right? That would be a pretty straightforward step. You don't really need to overly explain to AI. But some other things you're gonna want it to really follow, especially the more unique it is to you, follow the way you want it done.

19:25Rachel Woods back to the example of, a newsletter. for, let's say one of those steps is, create a subject line. the subject line, probably the way that you wanna write subject lines for your business is different than how I wanna write subject lines for my business, and some other people are out there putting, emojis and all caps and all that crazy stuff, right?

19:40Rachel Woods And so what I wanna do in my playbook is I wanna make sure that I am writing out some more specific instructions of how I want the AI to do that step and, there's a whole art, we can go into that. But, but really once you've done that, defined those four components and filled out some more instructions for the key steps, then that's it.

19:56Rachel Woods You've defined the work, and that's all AI needs to do to run it.

20:00Isar Meitis Awesome. So yeah. I will add two small cents, and they're both related to the same thing. One is examples. AI is really good at following examples, both of the inputs and the outputs and the process, so you can actually give it... and I'll say something that I really love doing recently because the, the last six months, the, the models became so good, they just figure it out.

20:18Isar Meitis So let's say you're doing... I just did a few workshops for, finance people, so it's a lot of, Excel stuff and going from this inputs to that output. And historically what I did is exactly what you're saying. Okay, we run this when, blah, blah, blah, blah, blah. here is how the input files look like, here's how the output files look like, and I would explain the process step by step.

20:35Isar Meitis I go to column I and I look at what's in column I in this sheet, and that is then compared to what's in column V in that sheet, and then we manage this based on whatever, and then we... And now I just give it to AI, I said, "Okay, here's the input, here's the output. Tell me if you can figure this out."

20:49Isar Meitis And it will sit there for five, 10, 20 minutes. It says, "Okay, here's what I think that's happening." And usually it gets you 85% there and you can just tell it what the gaps are and fill out the details. But examples just work very well, and AI is, A, very good at following these examples, and B, very good at- deducting based on these examples, I will say that.

21:08Isar Meitis And the second thing that I wanna say that I'm sure is a part of your process, you're never done with version one. So you build this once, and a lot of my processes, and I don't call them playbooks, but they're identical to what you're just saying, some of them are probably version 87. A, because I'm learning and I'm...

21:24Isar Meitis and I just wanna make it better. B is because things are changing in the environment. Microsoft comes up with a new version, and now Outlook looks differently, and it has another field that you can use or, stuff like that. So the environment is changing, the requirements are changing, and you, now that you have this, you always have new ideas, and I'm sure you're just like me.

21:40Isar Meitis You're like, "Ooh, what if this can also do that?" and then you go and add either a new playbook or call it whatever you wanna call it, or you fine-tune ones that you already have, to do this one extra thing that you want because now your morning routine includes one more thing. your yoga skill or whatever it is that you wanna do that you can add to that.

21:57Isar Meitis So Let's look at an example, like an actual one, and see how it's built and how it's done, and what is the inputs and outputs. I think that will make it very clear to people.

22:06Rachel Woods Awesome. So why don't I do this first? I am going to share my screen, and we're gonna go through, So as we're talking about these playbooks, you might be wondering, okay, are we, like, making these in Claude? Are we making these in ChatGPT? Like, what tool are we using? what we actually do is we have a philosophy we call own the playbook, rent the tech, which is, you should be creating your playbooks in a document that is outside of these tools.

22:33Rachel Woods That could be Notion, it could be Google Docs, it could be, like a lot of my team, we use GitHub. it-- But the reason is then you actually, focus on, the thinking and the operational process itself versus getting sucked too deeply into the tool. 'Cause these tools, like you're saying, like they s- they change all of the time, and if you're constantly having to change your process 'cause the tool changed, that's gonna suck, for lack of a better word.

22:57Rachel Woods but what's better is if you, again, create, the playbook to be a little bit more tool agnostic so that then... And I like to say a, a good playbook, literally a human could go through this and read it, and do it. That's really the sign of a, a good playbook. So let me just show you guys what a couple of those look like.

23:13Isar Meitis I'll say one thing while you're opening your screen. The- Yeah if you're looking more to know about what Rachel just said, a few episodes ago, because I'm not sure when it is actually going live. So when you're listening to this, look a few episodes back. There's an episode where I'm sharing how I'm working with Claude and ChatGPT at the same time completely seamlessly, and I can connect whatever other tool I wanted.

23:32Isar Meitis These are just two of my main tools, and how the whole document-based AI solution makes the tool completely irrelevant, and it gives you a lot of benefits beyond what Rachel said, that they're changing. You can manage your tokens across more than one tool, and then instead of maximizing and paying more, you can go, "Oh, I...

23:49Isar Meitis For the last week I'll switch to this," or just ChatGPT is better at these tasks and Claude is better at these tasks. I can assign different tasks to different, tools and so on. So yes, I agree with you 100%. The tool is completely irrelevant to how the whole thing should work.

24:04Rachel Woods Yeah. I love it. and that sounds like a valuable episode. I'll just plus one. Whenever that episode drops- everybody should go and listen to that. Yeah. I love it. okay. So let me, let me go through, let's see. I'll do a LinkedIn post writer playbook first, 'cause I think, this is one that people really love, and this is literally how I, part of my process for creating content which is, again, so zooming out, to get AI to do any process or any work, you just need to find those same four components, right?

24:31Rachel Woods The trigger, the inputs, the steps, and the outputs. And and this is like a little pretty template that we use just to be some extra scaffolding. Again, you could do this in Notion or anything. But if you look here for writing a LinkedIn post, what do I have? I have my trigger, which is that it, starts when you send a message, my, inputs, which is what are my raw thoughts or what I wanna talk about?

24:49Rachel Woods Then I have my steps, and so for this, playbook that's gonna help me draft LinkedIn posts, I want it to actually match the process that I use when I'm writing. And so what I do when, whenever I was thinking about this, it's not just like I just write it out of the blue. I don't know if people do that, but, I'm not that good.

25:05Rachel Woods I have to actually think about it, right? So what do I do? I take my brain dump, and I wanna, extract what am I actually trying to talk about? And that's, my one main point Then, what I, learned is if I can kinda, think about the audience who's, reading the post, and I can almost kinda predict what follow-up questions they may have, and I can actually try to answer those proactively in that content, then that post is gonna be better.

25:30Rachel Woods And so that's the second step in my process. Then whenever I'm-

25:34Isar Meitis Yeah, before, before you scroll down- Yeah just to, for those of you who are not seeing the screen, every one of these steps that Rachel is describing has a step input, which is in many cases probably the output of the previous steps or previous steps, depending on how far we go down that rabbit hole.

25:47Isar Meitis Then there's step instructions on exactly what to do in that step. Then there's a success criteria. then there's the response format, so again, what the output actually looks like. And then the actual outputs themselves. So th- these exist in every one of these steps as a very detailed definition on how to do that particular step.

26:05Isar Meitis So if you can see, it is very similar to the bigger process. Instead of a trigger, there's an input, but then there's the instructions, and then there's the formatting and stuff like that. So it's very similar. And something else that I have, two things that I have in my process, one that you mentioned in theory, but you address it when you give it the inputs yourself, and the other that you already have here, is who's the target audience and what's the success criteria?

26:28Isar Meitis So in my process that is very similar to your playbooking process, I have these built into the process. So it's always asking me, who's the target audience? Is it just me? Is it other people? Is it people that are not in my company? who are going to be using this? what's their preferences, level of knowledge, and stuff like that.

26:41Isar Meitis And then what's the success criteria? if this works, what happens?

26:45Rachel Woods Yeah. And, yeah, great call-out because back to what I was sharing, like again, I think when you're first getting into breaking down your work, especially if you are l- watching the screen share right now, it might be like, "Oh my God, that's a lot" to, that's a lot to define. Especially like we're usually not used to being this specific with what we want done, right?

27:04Rachel Woods oftentimes it's all this like knowledge that's just kinda "I don't know, it's just the way I do it." And you will, again like I shared, you'll eventually wanna get to the place where your instructions for each step really do look pretty detailed when those steps are unique to you. And if I were to just zoom out and say the easiest way to start, doing this, it's really just even, even just doing the, the steps themselves, right?

27:27Rachel Woods Of, thinking about when I do a LinkedIn post, I don't just write it. I extract the point, I ask myself follow-up questions, and then I, keep going from there. So there's layers to this, of course, right? and so yeah, I think that's probably, helpful. The other thing, maybe I'll just throw this in here, too, 'cause you might be thinking, "Wow, man, that's like...

27:43Rachel Woods Can I use AI to make these playbooks?" "Can I get the AI to, write it itself?" So in our work, what we do is we actually force people to write their first playbooks by hand, because if you haven't yet thought in this depth of level, it's gonna be really hard for you to what? To be the reviewer and actually understand what the AI did when AI writes its own playbooks.

28:05Rachel Woods Once you have this kind of way of work down, though, of course, you can use AI to speed up the process, right? We have a playbook that writes playbooks that we use, in our business. But, Okay yeah, but anyway, so to keep going down this, LinkedIn post, one, the, third step here is I have the AI help me reframe it.

28:23Rachel Woods the fourth step is I then have it draft the post. I can also, the way we do it, we, pull out things like examples or reference material or brand voice guides and stuff so that then also you can share them between playbooks. and then, edit it based on my style guide, and then the output is just LinkedIn post.

28:40Rachel Woods And so that's just one example. what I wanted to do is just show you, again, if there's, one takeaway I want you to really see when you look at this playbook structure, it's if I can just learn to do triggers, input steps, outputs, I can playbook anything Because that is truly the same structure for every single type of work.

28:58Rachel Woods And so wildly different use case away from content creation is actually one of my favorite playbooks, my project planning playbook. which basically, 'cause I, I love this, people oftentimes are like, I don't really do p- repetitive work and processes. All my stuff is, one-off." And I'm like, "Oh, so you have, a lot of projects, right?"

29:13Rachel Woods And people are like, "Yeah." And I'm like, "Great. Project planning is a process." And you know what? When I playbooked my project planning, I measured before and after playbooking this, and we now as a team do three times the number of projects at the velocity that we've accelerated because we've playbooked it, and so now AI's helping us get this work done.

29:32Rachel Woods But to playbook it, same things. So the, this one, the trigger is we just start it whenever we're ready to plan a new project. the inputs are project description. And then I just have the same thing of, the steps. We have the AI ask questions to clarify the project. We have the AI then brainstorm approaches, recommend one, and make sure that sounds good with us.

29:50Rachel Woods We have it then clarify the timeline and tasks, draft the project plan. And then my favorite thing, and it's hidden a little bit in here. I'm just gonna You can see. The instructions can get really detailed as you start playbooking this stuff. But is we have the AI, go through the project plan and say, I can actually do this task.

30:09Rachel Woods I can do this task. I can do this task," as it's, assigning work out, right? So it'll be like, "Okay, Christie should do this. Rachel should do this, but, I, as AI, can do this." So that then whenever the AI finishes helping us plan the project, it can basically be like, "Okay, I'm ready to start. Do you want me to start on my tasks?"

30:24Rachel Woods And then that way, oftentimes that means that a third or two-thirds or more of the project can be done before we even are, like, exiting out of the, chat with the AI. that's just an example again of a very different type of work, but same playbook structure,

30:37Isar Meitis I love it. what do

30:37Rachel Woods you say? should we run this? Yeah, go ahead.

30:39Isar Meitis Yeah. So you go start running this, and I will do... I wanna add my two cents to this that is very relevant to what you just, said, because again, we're very similar in our concepts and the way we approach things. So I have, a, a whole system that is connected to every meeting I finish. So every meeting I finish gets transcribed, it gets pulled immediately, and Claude reviews the whole thing with a very set, very specific set of requirements, and then it opens tasks for me in ClickUp, which is the same like Monday or Jira or Asana or whatever, Notion, whatever task management tool you're using.

31:08Isar Meitis So it's opened the task there, but then it runs through the tasks that it's opened and say, "Okay, which tasks do I know how to do?" And then it does all these tasks. so at the end of each, whenever I wanna jump into that board, I have the tasks that I need to do, tasks that other people need to do, they just get assigned, but then there's tasks that needs reviewed, and they're assigned to the right people to review because AI already did them.

31:30Isar Meitis And I have another board in which the AI collects the tasks that it doesn't know how to do but gets assigned a lot. So that's basically the next thing you wanna build, right? So if these tasks, it happens every single day and the AI doesn't know how to do them, that's probably the next thing you want AI to know how to do.

31:46Isar Meitis So then you will assign that to somebody to develop in your language, the playbooks for that. You develop the playbooks for that, and that nobody ever has to do this again, just to review the work, in the end. So very similar, mindset and concepts, and it becomes really fun to see how you need to do less tasks in everything that you're doing because AI learns how to do more and more things.

32:06Isar Meitis And then, like you're saying, just the velocity becomes bigger. It's not that you're doing less work. You're just doing more work with the same amount of people

32:12Rachel Woods Yeah, totally. I love that. We have a s- the same setup in ClickUp and everything, Awesome it's really the- Great

32:17Isar Meitis minds, right?

32:18Rachel Woods Yeah, it's really And back to like that human analogy, like that's why it works is because that's how, people have always managed delegation and tasks. And so again, if you're ever stuck with "How do I get AI to do this thing?" Just think about if you had people, like extra manpower, how would you set up the operations, and that's usually the path.

32:36Rachel Woods cool. Okay, so I opened up, Claude, and this is just a, little, I like to have my disclaimers, a little demo account so I don't, reveal all the, I've learned my lesson of all the, the secrets and things, Yeah but, but okay, once you've written the playbook, again, you have it, in your, in your document, and if you wanted a really easy way to save these inside Claude, for example, and the same thing works in ChatGPT, is you can go to customize, and you can go to skills, and, you can add, upload the skill or even chat with Claude and be like, "Here's my playbook," like download it as a document, and then have it turn that into a skill.

33:10Rachel Woods Because skills are a great place to save these playbooks. now, we could get into that like there's a lot of ways to build skills and structure skills and, the way I really think about when you're putting a playbook in a skill, it's because you're, saving that process. And the, the planning project playbook, what that looks like in Claude, as you'll see, it's just the exact same thing, the triggers, the inputs, the steps, et cetera.

33:33Rachel Woods And I've really just taken the, document that I wrote and put it in here. Now, once you've saved it, the nice thing is then, it can be accessed from, anywhere. You can have the AI kind of know when to pull it in, or you could even, do a slash command and have it, start that skill directly. And so we'll do, help me plan a, a wine night where we drink pinot noir and write playbooks That'll be my next- Am I

34:07Isar Meitis invited? that sounds like my kind of- my kind of event.

34:10Rachel Woods I've been thinking, you and I end up at a lot of the same, events and stuff. we should, do

34:15Isar Meitis a-

34:15Rachel Woods 100% a side thing.

34:17Isar Meitis 100%.

34:19Rachel Woods And so what it's doing here, and part of what, again, 'cause, 'cause we're operations n- nerds over here, right? Where we really wanna make sure it's following the process. So I actually have in the playbook itself, for it to tell me that it's following the steps so that I can, it improves the reliability and also improves that I can kinda audit it later, right?

34:36Rachel Woods Go look to make sure it actually did what I wanted. So this is what it's doing. It's already starting with step one of clarifying, the pitch. And so I can go through and, I'm not, reading those, but,

34:45Isar Meitis Again, those of you who are not watching the screen, it's asking Rachel a bunch of questions, and Rachel has to respond to these questions. those of you who haven't used Claude, or ChatGPT for that matter, it pops these, multiple choice questions kind of thing, and you can always type something else and tell it if it's not one of the choices, which I end up doing a lot.

35:01Isar Meitis But, combine that with voice typing, which I assume you do as well, makes it relatively easy. So even if it's not in one of the choices, you can always just click on something else and then voice type your answer and continue from there

35:11Rachel Woods Yes. And I, yeah, part of my, morning, walk is I even have, playbooks that I'm running in voice, so that's a great, another great call-out, yeah. so once it did that, then, it's now following step two of brainstorming approaches. So I can say confirm. Now, this is a playbook that's, The way it's written is it's designed to stop and ask me, so it's a little bit more of a collaborative playbook.

35:31Rachel Woods But you can also have these playbooks just go through and run all of these steps without your input. again, the beauty of learning just this way of structuring things is it works across everything, but anyways, so that's how, I would keep going through it,

35:42Isar Meitis and then- a- another thing here, by the way, that I saw that we're doing very similarly, when it's deciding what to do, it's giving options, it's defining the pros and cons of the options, and it's telling you its recommendation and why. And I tell a lot in my courses that I teach, I push back on at least 50%.

35:57Isar Meitis And when I say I push back, it doesn't mean I don't agree with the AI, it means I push back. I'm like, "Okay, I see you picked this, but what about this and that?" Or, "Did you consider this option?" "Ooh, this is an interesting one, and let me look at that." So I sometimes will go with the, what it selected, but if I have any doubts, I will push back and ask it for more options, or what about this, what about that?

36:17Isar Meitis And then I will decide whether I agree with the AI's recommendation or not.

36:20Rachel Woods Mm-hmm. Yeah, I love that. And then even, again, as you're, we kind of, like, zoom out and you're thinking about the way that you want to run your business, right? I think kind of an interesting level you'll reach once you start doing this is, at least in my experience, you realize there are some things that you always wanna have your input on, and there are other things where The risk or the cost of error is, like, pretty low.

36:45Rachel Woods And so then I always recommend, like, starting out with your playbooks being super collaborative, and then you'll learn where you can trust, and then those can become the ones that you set up, like we're talking about, with, like, ClickUp running them instead of you having to be, for example, in the chat running them.

37:02Rachel Woods So- Yeah I know we're kinda, coming up on this already being a long episode. There's one other thing, that I wanted to show, 'cause I think Yeah, I think we went through this playbook-

37:11Isar Meitis Yeah

37:11Rachel Woods well enough. Yeah. I

37:11Isar Meitis don't think we gotta- Okay we need to run it all the way to the end. People are probably really interested in the whole wine thing, but they can listen- to a different episode

37:18Rachel Woods on that. we'll keep you posted, yeah. But, and this is something that's, I like to just, again, like, um, be, super open and even kinda just sharing stuff as it's kinda newly emerging. But that progression that I'm talking about, of really thinking operationally about your work, and how you can keep scaling yourself, I think that really ends up being, like, a, such a big unlock.

37:39Rachel Woods And so this is a, little visualization of my current, kind of personal productivity system where I have, playbooks that are all running and I'm, I bucket them whenever they start stacking them together into almost like a subsystem or this becomes which, agent is running them and, I have, recently split out stuff more explicitly between what are the playbooks that I actually wanna collaborate with the AI on versus what are all the things that AI can just do in the background for me.

38:10Rachel Woods And then each of these have, escalation, processes, that then move into my, checking decisions log, which is where I need to be the one to, decide stuff. And so the way I think about this is, again, if I were, if I originally hired an assistant and I got them to do my email and my calendar and all these different things, and that was going really well, but now the assistant, is gonna either need to hire other assistants for it to go...

38:35Rachel Woods oftentimes, I remember when I was at Facebook, the, chief of staff, executive assistant, people, they had, their own assistants, right? 'Cause there's just so much work, that can get done. and so it's kinda that same thing with these playbooks, is thinking about how can you, codify the process so that then it doesn't have to be something that you're in the loop on at all.

38:53Rachel Woods And so yeah, so right now it's I have my overall chief of staff, which is the one that I work with, but then, it's all of these are running in the background of email, calendar. I have an EA and a projects, one that are kinda doing different things that don't need my input. And so each of these playbooks, back to the Lego blocks, that's really how you get there, is by starting one thing at a time and then being able to keep iterating and keep growing from there, I

39:14Isar Meitis love it. Again, very similar approaches to the stuff that I'm doing. and the interesting thing is it's not because we shared information, it's because that what makes sense when you run a business, right? it has nothing to do with AI. It has to do with being able to define operational processes, make sure they're running effectively and consistently, so both these things matter.

39:37Isar Meitis And then once it's there, then you can delegate more and more of that work, and you need to initially monitor and then even monitor less as things progress, and that's the whole biggest unlock, right? Is that people ask me, "Do you work less now?" I'm like, "I actually work more because the temptation to keep building more of these things."

39:58Isar Meitis So I work more hours than I did two years ago, but I do 10X the output. So I work 15% more hours, but I do 10X the work. And you multiply that by the amount of people in your organization, and you understand the implications of that on the revenue and the growth of the business. Rachel, this was absolutely fantastic.

40:19Isar Meitis Everything I hoped for and more. if people wanna follow loo- follow you, learn from you, work with you, what are the best ways to do that?

40:26Rachel Woods Yeah. This was so fun, and I love that, yeah, we're both, we're both, the ops nerdy type. I hope this was really helpful, for people. And if you wanna connect, I am really active on LinkedIn. That's probably the best, spot. But really on, on all social channels, I'm there hanging out. And then if you wanna learn more about our company, our URL is runonamp.com.

40:46Rachel Woods So we can, we can nerd out about anything, and I hope that after today you guys have, been inspired to write some playbooks. 'Cause remember that anything you're doing, you can playbook that, truly. So...

40:57Isar Meitis Amen. Thank you so much, Rachel. This was really, really fun. I'm sure people learned a lot. I learned a lot. Gave me a lot of small ideas like, ooh. and that's always why I love these conversations, like there's always these little tricky things like, ooh, I don't have this, or I should update that.

41:08Isar Meitis So I appreciate you. I appreciate you taking the time. Thank you so much.