71 | How Rachel Woods Uses AI Playbooks to Scale Operations

1 - > Audrey Chia: Hello, and welcome back to The AI Marketer's 2 - > Playbook, where we cover actionable frameworks to help 3 - > you leverage AI and marketing strategies in your business. 4 - > I'm Audrey Chia, your host, and today I have with me Rachel 5 - > Woods, CEO and co-founder of AMP, which stands for AI 6 - > Momentum Protocols, an AI operations education and 7 - > implementation firm helping businesses turn AI into an 8 - > actual teammate. 9 - > Now, Rachel has been working in AI for over eight years and 10 - > started her career as a research data scientist on Facebook's AI 11 - > team, even before most companies were talking seriously about AI.

12 - > Now, she's also the person who coined and developed what is now 13 - > known as AI operations, and has trained more than 500 AI 14 - > operators through her AI operator bootcamp. 15 - > Rachel has been featured in Forbes, on Tony Robbins, and 16 - > across major publications. 17 - > But what I love most about her approach is that she's really 18 - > technical, but also incredibly operational. 19 - > So Rachel, welcome to the show.

20 - > Rachel Woods: Thank you so much for having me. 21 - > It's so good to see you. 22 - > I'm so excited for this conversation. 23 - > It's gonna be a blast.

24 - > Audrey Chia: Awesome. 25 - > Can you tell us a bit more about, you know, your backstory? 26 - > Now, I know you were working in AI long before the current AI 27 - > wave. 28 - > What first pulled you into this space?

29 - > Rachel Woods: Yeah. 30 - > Um, you know, it's so funny, I feel like I was just a nerd, and 31 - > then I became a cool nerd because the world started caring 32 - > about what I was nerdy about, which is always convenient when 33 - > that happens, right? 34 - > Um, and, and so when I first got into AI, um, it was a little 35 - > bit, I wouldn't say by chance, but it was, I was mostly 36 - > obsessed with how technology and business kinda came together and 37 - > could solve problems, and I liked marketing a lot.

38 - > And so whenever the job came up to be able to work at Facebook 39 - > in their AI research team on the ads stack, you know, as someone 40 - > who likes those three things, it was just like,"Oh, my gosh, 41 - > dream job. 42 - > Of course I'm gonna do this." Little did I know how much I 43 - > would learn about, again, like, AI and its capabilities then, 44 - > um, not to mention how much the tech has progressed since. 45 - > Uh, one of the things that blows my mind is I was using language 46 - > models back in the day, but they weren't anything nearly as large 47 - > and impressive as they are now, um, and just how the world has, 48 - > has advanced.

49 - > Audrey Chia: Wow. 50 - > And what do you think is the biggest change you have seen, 51 - > you know, besides the advancement of it? 52 - > The way, has the way people use AI also changed along the way? 53 - > Rachel Woods: Oh, absolutely.

54 - > I mean, whenever I started my career in AI, you would really 55 - > have to have kind of a machine learning team, data scientists. 56 - > It was a pretty technical build-out to even start playing 57 - > with this stuff. 58 - > You'd also have to have all your own data centers and warehouses 59 - > and all this stuff to train the models. 60 - > And then maybe once you set all that stuff up, then you could 61 - > start experimenting and maybe get something kind of 62 - > interesting and useful.

63 - > And so you had to, there was so much of a startup cost to using 64 - > AI. 65 - > And then, I mean, now we can just be on our phones typing in 66 - > to chat, asking AI to do completely new and novel things 67 - > for us. 68 - > So just a complete 180 of how accessible it is. 69 - > Um, and whenever ChatGPT came out, that ended up being my, 70 - > like, oh, my gosh, this changes everything moment from an 71 - > accessibility standpoint.

72 - > Audrey Chia: Wow. 73 - > Um, I'm curious to know, Rachel, so you coined the term AI 74 - > operator, right? 75 - > Tell us more. 76 - > What exactly is it, for folks who may not know?

77 - > Rachel Woods: Yeah. 78 - > So I think that one thing that's really helped me is looking at 79 - > AI as really being a very big umbrella space, right? 80 - > There's tons of different ways you can use AI. 81 - > There's AI in the, you know, more creative space, especially 82 - > with, like, image generation and video.

83 - > Um, there's AI in the traditional sense with machine 84 - > learning models and all the algorithms and stuff. 85 - > And I, uh, from both the AI background and I've also been an 86 - > entrepreneur since the age of 15, um, I kinda fell in love 87 - > with this use of AI in business to help it, um, or help you 88 - > operate better, right? 89 - > To be able to think about what you want done, and then to 90 - > organize it in a way that then AI can help take that stuff off 91 - > your plate.

92 - > Um, and so whenever we first started using AI in this way, 93 - > you know, in the systems and processes in the business, there 94 - > really wasn't a word for it. 95 - > This was back in early 2023, and I remember sitting down with my 96 - > co-founder, who's my sister, and, um, our team at the time, 97 - > and we were throwing all sorts of words up on the wall- of, 98 - > like, should we call this, like, AI workflow implementation? 99 - > We're like,"No, that's, that's too specific."

Um, or should we 100 - > call it... 101 - > You know, there's just so many different types of words, and we 102 - > finally were like,"You know, this is just the new version of 103 - > operations," right? 104 - > Wow. 105 - > Which is AI operations.

106 - > Um, and so that's why we decided to start calling it that. 107 - > And then, of course, the hat that you wear when you're 108 - > someone who is using AI for AI operations naturally lends 109 - > itself to being the AI operator. 110 - > Um, and that's kinda how that all came to be, and it's been 111 - > fascinating to see then, you know, in just the last few 112 - > years, how widely that's been picked up, and I lo- I love it, 113 - > so. 114 - > Audrey Chia: Yeah.

115 - > I saw so many folks who call themselves operators, 116 - > go-to-market operators for AI or, like, fractional AI 117 - > operators. 118 - > I mean, it's very cool to know that you ladies are the one who 119 - > started it all- just from, you know, an idea in your head. 120 - > Rachel Woods: Yeah. 121 - > At least we, uh, we sure did argue over what was gonna be the 122 - > best term- that made sense for us.

123 - > So this is what, you know, just what we started calling it, so. 124 - > Audrey Chia: Yeah. 125 - > Awesome. 126 - > Yeah.

127 - > But can you tell us a little bit more? 128 - > When you say operations,'cause in a business, right, ev- so 129 - > many things, you know, can be supported with AI. 130 - > But what are some very tangible examples? 131 - > So let's say you go into a new business now.

132 - > What are some areas you look at? 133 - > Rachel Woods: Yeah. 134 - > So when I think about AI operations, it's really just all 135 - > of the systems and processes that power the business. 136 - > So it is across every department, from marketing to 137 - > sales to how you onboard clients or customers to the support you 138 - > get, even to how you operationalize maybe your 139 - > service so that you can efficiently deliver a service 140 - > all the way to your, uh, finance and, you know, HR or their 141 - > planning functions.

142 - > Operations really does touch all of those pieces. 143 - > Um, and so there are ways to look at the processes in each of 144 - > those departments and then see, oh, okay, like, we could either 145 - > have people do this process or how could we set it up so that 146 - > AI can do it. 147 - > Um, and that's what we spend our time on is, like, how do you 148 - > take a process that was being done by hand and now set it up 149 - > so that AI can do it reliably? 150 - > Uh, and our method that we use, that we, we call it, is our 151 - > playbooking method.

152 - > So it's kind of funny'cause I'm on the Playbook, uh, you know, 153 - > podcast and, um, obvi- obviously it's a word that's, uh, near and 154 - > dear to both of our hearts, but that's what we call, um, 155 - > whenever we take a process and turn it into something that AI 156 - > can do. 157 - > Audrey Chia: Wow, interesting. 158 - > Given that there are so many functions where AI can support, 159 - > right, how does a company even decide where do I begin? 160 - > 'Cause technically they have so many things they can do, 161 - > everything is possible, but at the same time they might have to 162 - > figure out where to start and I'm sure there could be some 163 - > hesitation, uh, especially initially when they don't know 164 - > what's gonna happen and whether it's gonna work or not.

165 - > So where do companies begin? 166 - > Rachel Woods: So there's two ways to think about where to 167 - > start. 168 - > One is really focusing on the word operations, right? 169 - > Which is if we just took AI out of the equation, where would you 170 - > invest in your operations?

171 - > And you'd probably do it in the stuff that's gonna have the 172 - > highest impact or, you know, if we solve this we would have a 173 - > breakthrough in the business and be able to do more or grow more. 174 - > Those are the same types of areas that usually have really 175 - > high impact AI operations projects, um, because it's, it's 176 - > really the same thing, just a new flavor. 177 - > Um, but maybe my more counterintuitive advice and 178 - > observation from now doing this, this specific work for three 179 - > years has been, um, that actually if you try to pick the 180 - > best use case as your very first use case, you are 181 - > overcomplicating what you need to actually be focused on and 182 - > you're gonna go slower than if you instead just prioritize 183 - > what's gonna be the fastest way for us to get momentum and for 184 - > us to learn.

185 - > And so when we work with teams, we say,"Amazing. 186 - > So glad that you have really high impact stuff that you wanna 187 - > use AI for, and l- great, let's put that second on the roadmap." 188 - > The first thing we should do, though, is this is a whole new 189 - > capability and mindset and way to think for- Mm a team of how 190 - > to use AI, how to be the one defining what the AI does, how 191 - > to check the AI's work. 192 - > And so let's actually just start with the easiest possible, you 193 - > know, use case or s- you know, something that's still gonna be 194 - > useful, but, uh, that gets the team momentum, and that way then 195 - > everybody can learn, and then your second project goes so much 196 - > better.

197 - > Audrey Chia: Oh, that's a very interesting perspective. 198 - > I think people are always looking for, like, an ROI first 199 - > kind of project, but like what you're saying- Totally starting 200 - > from the foundations. 201 - > so Rachel, you talked a bit about the team, right? 202 - > I'm curious to know, for a team, right, a team member, how do you 203 - > actually teach them or upskill them so that they are 204 - > comfortable using AI and they know when to use it?

205 - > Rachel Woods: Yeah. 206 - > So what we've found is it's best to think about, again, this, 207 - > like, AI operations bucket more as a capability that's gonna be 208 - > across all sorts of stuff that you're doing. 209 - > Audrey Chia: Mm-hmm. 210 - > Rachel Woods: And so it's kind of similar to the first time you 211 - > learn to, like, write down a process or make a checklist.

212 - > Yes. 213 - > You wanna give people the opportunity to playbook or 214 - > create a process AI could follow, um, for even just a 215 - > small piece of their work. 216 - > And once they have that first win, then it clicks, and then 217 - > you can start using that momentum to do more and more 218 - > things. 219 - > Audrey Chia: Hmm.

220 - > Interesting. 221 - > And do you feel like a lot of team members are also resistant, 222 - > or are they pretty open to adopting AI? 223 - > Rachel Woods: It varies, and what we found is the biggest 224 - > indicator is not actually the person, it's more about the 225 - > culture of the company. 226 - > Audrey Chia: Oh, how 227 - > Rachel Woods: so?

228 - > So I'll give you an example. 229 - > Mm. 230 - > So if you're, um, if you're somebody that is maybe hesitant 231 - > about AI, but you work in a team where it's very clear that 232 - > there's more than enough work, more than enough demand for what 233 - > the company does, and so okay, like I could see how if AI did 234 - > some of these things I'm doing, there's tons of stuff that's c- 235 - > now I can do instead. 236 - > It's a lot easier for you to then lean into and trust that 237 - > process versus, you know, frankly, we do work with teams 238 - > where I think it's very valid if you're sitting there and you're 239 - > like,"Well, if AI's gonna do half of my job, but we don't 240 - > have more work for me to do, then what?"

You know? 241 - > And so, um, I think it's, yeah, like those are bigger culture 242 - > and business viability questions. 243 - > Um, but if you have the demand and a really strong business, 244 - > giving people and painting that path for them, uh, we've seen 245 - > people really open up. 246 - > Audrey Chia: Interesting.

247 - > I also wanted to talk about maybe a slightly more sensitive 248 - > topic, but I think it's also important to discuss. 249 - > So in Singapore recently, we had a wave of, uh, layoffs from tech 250 - > companies without a lot of warning. 251 - > Um, some people found that they had no jobs the next day, and 252 - > there was also like a sweep, you know, a sweeping statement their 253 - > companies made about, uh, p- them using AI, hence some 254 - > functions were no longer, um, as necessary as, as before.

255 - > So I was curious to know, so what are your thoughts on where, 256 - > where does the pe- the person come in when AI can take over 257 - > maybe not everything, but lots of functions that, like what you 258 - > said, people used to do? 259 - > Rachel Woods: So with any thing you're gonna have AI take over, 260 - > if you wanna trust that it's doing a good job, you need 261 - > somebody that has the judgment to define what good looks like. 262 - > And that really is what we've seen as like the role that has 263 - > to be human, has to be based on real expertise, has to be based 264 - > on, um, the background experience that someone has.

265 - > Like, that's how you have the judgment to even know whether 266 - > the AI is doing a good job or not. 267 - > So I think what's really challenging right now is there 268 - > are a lot of companies that, uh, they're having to shift around 269 - > their workforce because they, um, you know, are bringing in AI 270 - > to some areas where then, you know, one person can have a 271 - > judgment and that actually covers 20 or 100 previous roles. 272 - > Mm. 273 - > Other places, though, it's You really have a closer, like, 274 - > one-to-one relationship, right?

275 - > Yeah. 276 - > Where it's like if one person's doing something, and now AI 277 - > comes in and helps with that, great. 278 - > That one person can now do more. 279 - > Audrey Chia: Yeah.

280 - > Rachel Woods: Um, so I think if I You know, when I talk to 281 - > anybody about how they can position themselves, I really 282 - > ask, like,"What do you feel like you're really an expert in? 283 - > Where do you have judgment?" Mm. 284 - > "Where are you excited to build more judgment and more expertise 285 - > over your career?"

That should be where you're investing your 286 - > time, and then learning how to be the AI operator that's 287 - > running the AI that helps you, um, amplify what you do. 288 - > Audrey Chia: Definitely. 289 - > I think also for me, um, having that mindset of, okay, what is 290 - > Like what you've mentioned, the higher level skill set, uh, that 291 - > can continuously value add to the team. 292 - > Um, what are, what are the human skill sets that you definitely 293 - > need that cannot be replaced?

294 - > Perhaps even, like, relationship management, right? 295 - > Uh, these are things that are one step above, uh, that you 296 - > still need that human element. 297 - > Then number three, learning how to adopt AI or be comfortable 298 - > with it so that, um, when AI processes come into play, you 299 - > don't feel too overwhelmed. 300 - > Now, Rachel, going back to what you do, right?

301 - > So you talked a bit about the playbook. 302 - > Can you tell us just a snippet of what you cover in the 303 - > playbook and, and how companies can go about their own 304 - > processes? 305 - > Rachel Woods: Yeah. 306 - > So in our world, playbooking is really a method for how to take 307 - > a process and turn it into a process that AI can do.

308 - > You can give these, um It's really just a document that 309 - > outlines the step-by-step of that process, and you can give 310 - > that document to an AI in a chat tool. 311 - > You could give that document to an AI agent, and you can even 312 - > have systems of playbooks that are run by systems of agents. 313 - > Um, and so it's almost this, like, building block of defining 314 - > what you want the AI to do and building your operations, uh, 315 - > through writing these playbooks.

316 - > Audrey Chia: Mm. 317 - > And when companies first came to you versus now, right, that has 318 - > been, like, three to four years, do you think there has been any 319 - > change? 320 - > Are they more ready compared to before? 321 - > Rachel Woods: Oh, completely.

322 - > Yeah. 323 - > And also, the, the tech has changed a lot. 324 - > So in the earlier days of even ChatGPT, when you started trying 325 - > to have AI do more complex stuff, usually you had to go 326 - > into automation tools, right? 327 - > Like a Zapier or an n8n or, like, some of these more 328 - > technical tools, and it just took longer to set all these 329 - > things up.

330 - > Um, you know, I'm, I'm curious if you're using, like, Claude 331 - > Skills and, uh, some of these, uh, things that have come onto 332 - > the scene in, in the last six months. 333 - > I, I'm sure you're using it, uh, to the Nth degree. 334 - > But, um, it's so much easier to set these things up now. 335 - > The hard part is deciding what the process is in the first 336 - > place.

337 - > Um- Mm so I think not only the companies are more eager, but 338 - > the tech also is so much easier now- Yeah which I think is 339 - > really great. 340 - > Audrey Chia: Yeah. 341 - > It has made it so much more accessible for non-technical 342 - > folks, and I think one of the other, um, uh, uh, someone I was 343 - > speaking to recently said,"You don't have to be technical at 344 - > all, Audrey. 345 - > You just gotta go for it and then see what you manage to 346 - > build with AI."

347 - > Rachel Woods: Yeah, I think that the real hard part and the skill 348 - > is the systems and process thinking, right? 349 - > Um, and so you don't have to code anymore to do any of the 350 - > stuff the AI does, just build itself if you can think in the 351 - > systems and process, you know, so it stays organized. 352 - > Audrey Chia: Yeah. 353 - > Let, let's unpack that a bit, right.

354 - > So you mentioned systems and process. 355 - > So even for me, like, if I were to prompt for a copy, I break it 356 - > down into, like, seven to eight steps. 357 - > So that's my own internal system of how I break things down. 358 - > Can you give us an example of how does a business take one use 359 - > case and break down their process?

360 - > What does that look like? 361 - > Rachel Woods: So just like you're describing, um, I think 362 - > we as people tend to skip a bunch of steps when we think 363 - > about the work we do. 364 - > And so usually, um, what we find is when you really sit down and 365 - > force people to be like,"Okay, but then how do you do that? 366 - > Okay, that actually sounds like three steps right there."

And 367 - > you, you kind of pull apart what is the unspoken logic that might 368 - > exist in a process. 369 - > Um, there are a lot of ways to do this. 370 - > One of the, you know, most straightforward ways is to make 371 - > a process map if something is more complicated. 372 - > Um, the other way is to really just say,"Hey, Audrey, if you 373 - > had to teach me how to write copy like you do, what would be 374 - > your best effort and way of breaking that down in a way that 375 - > I could follow?"

Okay. 376 - > Mm. 377 - > That's probably the start of what the playbook could look 378 - > like. 379 - > Um, and then one of the best things I really like about 380 - > getting to make these processes that AI runs is it's really fast 381 - > to iterate, right?

382 - > So- Yeah um, you know, you could think, okay, here are the 383 - > different steps, and start running it, and then you're 384 - > like,"Oh, actually, step four might need to switch to step 385 - > five." And so you, uh, have a really fast feedback loop to get 386 - > better about your process. 387 - > Audrey Chia: Oh, interesting. 388 - > Um, I...

389 - > Now, when you mentioned it, I also thought about, you know, my 390 - > own kind of workflows. 391 - > And sometimes I realize that as a copywriter in the past, I 392 - > would used to come up with copy, just, you know, skip all the 393 - > micro steps in my head, right? 394 - > But then when I had to prompt for AI, uh, with AI, I had to 395 - > teach it almost how I came to that landing page copy. 396 - > So- Totally then figuring out what are all the steps before.

397 - > And I realized that it actually took a very structured way of 398 - > thinking, um, that I think not, that doesn't come as naturally 399 - > to everyone. 400 - > So it's almost like training yourself to view things in a 401 - > different lens so that AI understands you from that lens. 402 - > Rachel Woods: I completely agree. 403 - > It's, um, I think I've seen it really most similar to people's 404 - > ability to delegate, right?

405 - > Mm-hmm. 406 - > Like delegating to somebody else is actually really hard. 407 - > Yeah. 408 - > Uh, you don't like wake up, you know, or you're not born with 409 - > the ability to delegate.

410 - > Usually you have to like learn from somebody that was showing 411 - > you stuff, or like you try delegating, it doesn't quite 412 - > work, and you kind of build your own way of how to, um, empower 413 - > somebody else to do, you know, work the way that you want or 414 - > need it done, um, or the outcome you need. 415 - > And so it's, it's just like that, but for AI, you know? 416 - > You have to learn it. 417 - > Audrey Chia: Yes.

418 - > Yeah. 419 - > And the more specific you are, um, with your micro steps or 420 - > instructions, then I guess the more AI understands what you 421 - > want. 422 - > So I think also being able to build this ability to break down 423 - > your own thought process, see your, your work in systems, um, 424 - > is very important. 425 - > I've learned now that I can no longer just be a creative.

426 - > I need to be a creative person and a structured person at once. 427 - > Um, that's where AI comes in. 428 - > So Rachel, having said all of those things, would you be able 429 - > to share what it actually looks like? 430 - > Do you have a kind of demo or something that we can share with 431 - > the audience so they know what it looks like in practice?

432 - > Rachel Woods: Yeah. 433 - > Um, I have one that I pulled up. 434 - > So if I were to, uh, zoom out for a second, so what, you know, 435 - > we do is, uh, we take processes and we turn them into AI 436 - > playbooks that AI will run. 437 - > Um, what that looks like is usually we're writing that 438 - > process out first in a Google document or like just completely 439 - > outside of the tool we're using.

440 - > And the reason is, that gives us more space to actually think 441 - > about, like you're saying, how we wanna break up the work into 442 - > the smaller tasks. 443 - > Then once we create that process, and I'll show you the 444 - > structure in a second, then we have this philosophy, we call it 445 - > own the playbook, rent the tech. 446 - > So you can actually run that playbook that you wrote in any 447 - > tool. 448 - > So you could run it in Claude Skills, you could run it in, uh, 449 - > ChatGPT, custom GPTs.

450 - > Um, you could run it in like Microsoft Copilot agents if you 451 - > want to. 452 - > Ooh. 453 - > Like really any, any tool out there that can follow 454 - > instructions, you could run a playbook in it. 455 - > Um, and what's nice about that is then you're actually building 456 - > like your operations as an asset that you own instead of these 457 - > tools.

458 - > So I pulled up, um, like I, you can playbook anything, so I 459 - > could go through so many different examples, but I wanted 460 - > to pull up, um, one of my unexpected favorites. 461 - > Audrey Chia: Ooh. 462 - > Rachel Woods: Which is my project planning playbook And 463 - > the reason that it's unexpected is'cause you might think project 464 - > planning, that's not sexy. 465 - > Like- Mm we're not gonna do, we're not gonna do like, you 466 - > know, a marketing, some marketing thing or making an 467 - > analysis or report or research or a contract review or 468 - > something.

469 - > But, but actually, um, once I created this playbook, I 470 - > realized, wow, like I plan projects all the time. 471 - > And you know what? 472 - > If you actually turn project planning into a process, then, 473 - > uh, not only... 474 - > For us, we saw like our projects, we were planning them 475 - > better, which means they go smoother.

476 - > You don't have like, you know, down the line, oh, we didn't 477 - > think through blah, blah, blah, or we didn't decide who's gonna 478 - > own X, Y, Z. 479 - > Uh, and then I'll show you one of the coolest things is this 480 - > actually sets it up so that, uh, then the AI can even help you 481 - > execute the project more easily. 482 - > Wow. 483 - > So- 484 - > Audrey Chia: Let, let's see it.

485 - > Rachel Woods: Yeah. 486 - > So this is a fun one. 487 - > Um, so let me see if I can screen share. 488 - > All right.

489 - > Can you see this in Claude? 490 - > Yes, 491 - > Audrey Chia: perfect. 492 - > Rachel Woods: Okay, great. 493 - > This is a little demo account I have.

494 - > So, um, uh, this is not the full glory, I will say, of the 495 - > project planning playbook that I used, uh, I guess I probably 496 - > used it twice today. 497 - > Um, but this will give you a good idea of, of what I'm 498 - > talking about. 499 - > So, so in Claude, in, uh, Skills, um, what I've done is 500 - > I've uploaded my project planning playbook, and really 501 - > again, a playbook is a process. 502 - > And so for it to be a playbook, um, we say it needs four 503 - > components.

504 - > It needs to have a trigger, which is what starts the 505 - > playbook. 506 - > It then needs to have inputs, so what is the AI gonna ask for 507 - > before it starts running the playbook? 508 - > And then it needs to have steps. 509 - > Um, and so in here I have step one, step two, et cetera.

510 - > And then the very last thing is it needs, um, outputs And then 511 - > inside here for each of the steps, so let's go to this one, 512 - > uh, we have a lot of very specific instructions and I'm 513 - > probably not on time to go into all this today, but, uh, 514 - > basically kinda like you're saying, just the more specific 515 - > you are, and we have some ways that you can get really specific 516 - > to make sure that it follows these instructions. 517 - > Audrey Chia: Mm.

518 - > Rachel Woods: And so if you, um, I know, I mean, I'm sure you've 519 - > seen tons of different skills, setups in your day, right? 520 - > I would describe this as like, it's like the playbooking method 521 - > helps you write a skill that's structured as a process versus 522 - > there are a lot of other ways to build skills, right? 523 - > Different, different structures. 524 - > Audrey Chia: Can you explain that?

525 - > What do you mean by helping you to build skills as a, like, 526 - > process? 527 - > Rachel Woods: Yeah. 528 - > So, when you, um... 529 - > I don't think I have any here of, like, non-playbook skills.

530 - > But a skill really is just saving instructions in your 531 - > Claude account that then the AI r- can use next time you're 532 - > doing that type of task. 533 - > Yes. 534 - > And so it could be as easy, or, uh, an example of another type 535 - > of skill would be if you wanted, like, a brand guideline skill 536 - > that had all your brand guidelines. 537 - > That doesn't need to be structured as a step-by-step 538 - > process.

539 - > It'd be a little more like a reference guide or doc, right? 540 - > Versus this is a skill structured as, like, a playbook, 541 - > where it's very specific of, like, that, you know, you want 542 - > that step-by-step process. 543 - > Does that make sense? 544 - > I 545 - > Audrey Chia: see.

546 - > So when you're structuring your playbook, basically it needs to 547 - > wait for step one to be done before it goes to step two, so 548 - > then it forces it down a very systematic kind of funnel. 549 - > Rachel Woods: Correct. 550 - > Okay. 551 - > So then it's more reliable, and you can actually trust it to do, 552 - > uh, more in-depth work.

553 - > Audrey Chia: Got it. 554 - > Yeah, I think also- Mm-hmm 'cause, uh, when you mentioned a 555 - > brand guide, I also have a brand guide skill. 556 - > So, uh, in that brand guide skill, for listeners who may be 557 - > new to it, I put in information like my company colors, fonts, 558 - > logos, um, X, Y, Z, right? 559 - > So then it knows how to build a document that looks like Close 560 - > with Copy's document.

561 - > Um, in the same way, like, in that case, I'm giving it a ton 562 - > of knowledge, but I don't have to necessarily get it to follow 563 - > each step because it's just drawing from that pool of 564 - > knowledge. 565 - > Whereas I think what Rachel is also sharing is a very 566 - > systematic way of thinking'cause it's forcing it to go through 567 - > each step and even you have things like wait till you 568 - > confirm step one, right, before you move to step two. 569 - > Um, I think- Yeah that helps the process as well.

570 - > Rachel Woods: Yeah, totally. 571 - > It's like almost the skill you're describing is, like, a 572 - > context skill, right? 573 - > Um, versus this, yeah, is the process-oriented. 574 - > So tons of use cases for skills, right?

575 - > Um, and I think that's actually... 576 - > All these AI tools have tons of different ways to use them, and 577 - > so it's about, you know, learning the methods that work 578 - > for you. 579 - > But, um- Awesome. 580 - > So 581 - > Audrey Chia: how does this playbook play out?

582 - > Rachel Woods: Yes. 583 - > So this playbook, uh, so once I have it saved, and I'll, I'll 584 - > just run it as our way to see what the steps are. 585 - > Um, I can say,"Help me plan," and I can just p- uh... 586 - > If you do slash, you can actually search for the skill.

587 - > Um, and then I will say, uh, let's start there. 588 - > So what it's gonna do is it's gonna go read that skill 589 - > Audrey Chia: So while we're waiting for the skill to load, I 590 - > wanted to ask Rachel, how long does it take for you to 591 - > personally develop a skill? 592 - > Rachel Woods: It's a good question. 593 - > Um, we, and how I make playbooks is I outline them by hand, and 594 - > then I use AI to fill in the instructions, but then I do 595 - > still go through and, um, review and edit and, like, make sure 596 - > that it works'cause I want the process to be something that, 597 - > like, I trust.

598 - > So it's pretty fast, but I will say it's, it's probably a little 599 - > bit slower than a lot of people that, you know, are like,"Oh, I 600 - > just want AI to generate everything for me." Yeah. 601 - > Audrey Chia: Yeah. 602 - > Rachel Woods: So, um, so yeah.

603 - > Okay, so load up the skill, and then what it's doing is, as you 604 - > can see, it's already following step one. 605 - > Um, so for this one, it's saying,"Okay, well, I don't 606 - > really... 607 - > You know, I don't have the project that you're trying to do 608 - > at all, so you need to share all this context," um- Mm which is 609 - > great. 610 - > So I'll say, okay, so the context, um, we're working on, 611 - > uh, let's see, recording an awesome podcast with Audrey, um, 612 - > and want to, uh, make a cool demo.

613 - > Goal: explain playbooking. 614 - > Constraints... 615 - > C- of course, you can't type when you're trying to, but 616 - > that's okay. 617 - > AI can understand- 618 - > Audrey Chia: Yeah 619 - > Rachel Woods: typos.

620 - > Um, how it worked. 621 - > Uh- So let's do that. 622 - > And so what this is, and I wanna pause on this while that's 623 - > running, um, is it's following my first step, which if I go in 624 - > here, is I told it like, okay, a vague pitch, which is what I use 625 - > to describe the start of a project, um, produces a vague 626 - > plan. 627 - > And so it's gonna ask me for, uh, any information across the 628 - > context, goal, constraints, and succ-success criteria.

629 - > Yeah. 630 - > Um, ask with clarifying questions, and if still unclear, 631 - > ask more focused batch. 632 - > And so you can see now that it did that. 633 - > Um, let me see, make assumptions 634 - > Audrey Chia: I like that Claude asks a lot of clarifying 635 - > questions, and I realize that Claude asks better questions 636 - > than ChatGPT.

637 - > ChatGPT doesn't typically ask questions to you, just,"Okay, 638 - > run." But Claude does ask you a little bit more. 639 - > Rachel Woods: You know, that's funny because, um, I've observed 640 - > the same thing, and then whenever you're playbooking, you 641 - > wanna think about, like, do I want it to ask questions? 642 - > Audrey Chia: Mm.

643 - > Rachel Woods: Um, and so actually for a playbook, so we 644 - > say, like, if you write a good playbook, it should run the same 645 - > in any tool. 646 - > Um, but if you run that playbook in Claude, you don't specify, 647 - > it's definitely gonna ask you follow-up questions, right? 648 - > So, uh, if you don't want that, just putting that in your 649 - > playbook, don't ask any follow-up questions, you know. 650 - > Like, for this one, it, I think it would've asked more follow-up 651 - > questions, but I had said in here,"Do not proceed past two 652 - > rounds," right?

653 - > Audrey Chia: Mm. 654 - > Rachel Woods: Um, and so then now it's going to step two, so 655 - > brainstorm approaches. 656 - > Uh, I used an approach called the, I call it the, uh, the 657 - > 1-3-1, which is come up with three options and recommend one, 658 - > uh, end result, and so that's what it's done here. 659 - > Um, sounds good.

660 - > And so you can see it's following my playbook. 661 - > So step three of this one is to determine the timeline and 662 - > tasks. 663 - > Audrey Chia: Wow. 664 - > Rachel Woods: When does the pro- progress record?

665 - > Today. 666 - > What playbook are you generating? 667 - > Um, I'm gonna actually say LinkedIn post playbook. 668 - > And then, uh, nope And so that's doing step four And it's making 669 - > the task list.

670 - > Um, I can say it looks good And then step five. 671 - > So in our project planning process- Mm-hmm before we were 672 - > using AI, you know, just stop here, right? 673 - > Okay, made a, made a plan. 674 - > Audrey Chia: Yeah.

675 - > Rachel Woods: But what's great is because AI's helping me plan 676 - > the project, I could add another step, which is like,"Hey, AI, 677 - > you know what you can do." 678 - > Audrey Chia: Yeah. 679 - > Rachel Woods: Can you tell me what tasks you could actually 680 - > start on?" Yeah.

681 - > Wow. 682 - > Um, and then I can say like, "Yep, looks good." And then, um, 683 - > I have it make a final plan, and then I believe I had this one 684 - > also sending into ClickUp. 685 - > Um, yeah, so I don't know if my demo board is connected, but- 686 - > Audrey Chia: Got it.

687 - > So you also a- add additional connector so that then you can 688 - > ship it to, like, a ClickUp or any other platform, um- Yeah and 689 - > then you can, you can review it there. 690 - > Rachel Woods: Exactly. 691 - > And then for playbooks, what's great is, like, it's just as 692 - > easy as Let's find that step. 693 - > Add to ClickUp.

694 - > Um, it's just as easy as saying, like, you know, use ClickUp. 695 - > Now, if you were using a custom tool, you'd wanna put the custom 696 - > tool name in here. 697 - > Yes. 698 - > But, uh, yeah, AI's pretty smart, so.

699 - > Audrey Chia: Wow. 700 - > Rachel Woods: So, yeah. 701 - > That's so 702 - > Audrey Chia: cool. 703 - > Rachel Woods: Um, so yeah, now it's Oh, it's cr- trying to 704 - > create it in my actual board.

705 - > So we'll pause there. 706 - > Wow. 707 - > But, um, but yeah, so that's a playbook. 708 - > And again, this is a really simple one for something that 709 - > every single person plans projects, right?

710 - > Yeah. 711 - > And so what we've seen is, like, this, once we started using this 712 - > as a team, we were able to- Yes we measured it. 713 - > We started doing three times the number of projects- Wow because 714 - > they're just planned better, and then AI does, like, half of it 715 - > before you even start. 716 - > Yeah.

717 - > Um, and this is the exact same approach you can do for, like, 718 - > contract review for, uh, potential brand partnerships, 719 - > for example. 720 - > Mm. 721 - > Um, or writing your newsletter, or, uh, running and pulling, 722 - > like I just did a webinar this morning, um, for our community, 723 - > and we have a playbook that pulls all of the webinar 724 - > registration data, who showed up, what questions were asked in 725 - > the chat, pulls it versus the transcript, produces this really 726 - > lovely analysis of, like, what are some opportunities to 727 - > improve the webinar for next time.

728 - > Again, all a playbook, all in this exact same structure. 729 - > So, um, it's very- Very 730 - > Audrey Chia: super cool 731 - > Rachel Woods: very powerful. 732 - > Yeah. 733 - > Audrey Chia: Yeah.

734 - > I think also the possibilities are endless, right? 735 - > You just showed one use case, but I'm sure if you are 736 - > listening, you probably have so many other amazing use cases, 737 - > um, that you can already start thinking of, even as you watch 738 - > Rachel's demo. 739 - > Now, Rachel, you said this playbook can be applied across 740 - > different, uh, tools, right? 741 - > Do you have a personal favorite?

742 - > Rachel Woods: I, I'd say it's it's tough. 743 - > The tools, especially ChatGPT and Claude, I kind of view them 744 - > as, like, really competitive sisters. 745 - > Mm. 746 - > Like, they're just always gonna be kinda outdoing each other a 747 - > little bit.

748 - > So I actually also really like that I can move my playbooks 749 - > across these tools and, um, that's a, it's a huge benefit. 750 - > Uh, I will say our company does run off of Claude. 751 - > Um, so it does help in a team to have one tool and, uh, and 752 - > Claude's, Claude's quite good and they have a lot of, again, 753 - > like the connectors and all those features and stuff. 754 - > But yeah, I, I like them both.

755 - > Audrey Chia: I, I started with ChatGPT. 756 - > Um, that was my first love. 757 - > Then Claude came into the picture, then I was like,"Hmm, 758 - > so should I try a new lover?" Then I moved on to Claude, but 759 - > then I realized that I still wanted to, like yourself, keep 760 - > ChatGPT'cause you never know, right?

761 - > There could be something new that comes out, um, and it's 762 - > still great at certain functions. 763 - > So having both tools, I think also knowing when to use each 764 - > tool, uh, that also helps in the process. 765 - > For me also, for writing, uh, especially, I would recommend 766 - > using Claude, um, and also really testing out, uh, what 767 - > Rachel already demoed, the skills function,'cause it's so 768 - > powerful when you start thinking in skills, not just in minor 769 - > copy tweaks.

770 - > So Rachel, what do you think the future of operations could look 771 - > like for most companies? 772 - > Rachel Woods: I mean, I, I think it's gonna look like this. 773 - > Like, I think we're gonna spend a lot more of our time thinking 774 - > about how the work should be done rather than doing it. 775 - > Um, I'll actually give you, like, a wild example from last 776 - > week.

777 - > So I was, I was pretty sick last week, and, um, you know, being 778 - > out sick one day, o- okay. 779 - > But the second day, I'm like, "Man, like, I, I have gotta get, 780 - > uh, stuff done." Well, I went and I opened up my laptop and I 781 - > clicked the button on my computer that I- lets me, like, 782 - > talk to my computer, and I just told my AI,"Okay, run my 783 - > todaying, uh, playbook," which is the playbook that runs my 784 - > morning routine. 785 - > And I have that playbook actually connected to, I think, 786 - > like, seven other playbooks that do all my sales tasks, that will 787 - > follow up on proposals, that will do, um, marketing reporting 788 - > and that's scheduling meeting, like, all these different 789 - > things.

790 - > Um, and you know, my AI ran for quite some time, but then, uh, 791 - > once it finished, I was like, "Wow, this is actually, like I 792 - > was pretty productive today from bed of clicking one button and 793 - > saying,'Run my todaying routine.'" You know, and, um, 794 - > I've worked to set up the system that then makes that possible, 795 - > but it- all it is is playbooks and making sure it has access to 796 - > the right data and tools to be able to do the work. 797 - > And so I think, you know, you can imagine, like, you go to the 798 - > beach, uh, and you are e- enjoying your day and it's like, 799 - > okay, well let me just kick off this, this one thing that 800 - > creates then now 20 things that the AI is doing for me, but 801 - > following, you know, the process or playbook that I've already 802 - > outlined.

803 - > So, yeah. 804 - > Wow. 805 - > I think that's how companies are w- will operate, and it's 806 - > already how we're operating as a team. 807 - > Um, it's what we teach businesses that we work with 808 - > and, uh, how to operate.

809 - > So I, I just think that's the future, and the future's here. 810 - > Audrey Chia: Wow. 811 - > And it's like, you know, um, Tim Ferriss' Four Hour Workweek, I 812 - > think right now it's gonna be very, very possible for most 813 - > people to, to really start building out systems to help 814 - > them get, get things done. 815 - > Um, and Rachel, for yourself, how, how do you think companies 816 - > need to evolve and adapt?

817 - > 'Cause I do know, at least in my region, although companies are 818 - > saying they want to be, uh, open to AI, I, I sense a lot of 819 - > resistance because it takes up so much time, manpower, 820 - > bandwidth to, and resources to even think about 821 - > operationalizing certain parts of your business. 822 - > So what do you have to say to business owners who are still on 823 - > the fence? 824 - > Rachel Woods: I think it's a tough place to be in because the 825 - > reason you're on the fence is because you probably have a lot 826 - > of other stuff on your plate, right?

827 - > It's not like you have nothing to do, and so you're sitting 828 - > there going,"No, we still don't wanna do the AI thing," right? 829 - > You have other priorities. 830 - > I think what's challenging is how long are you willing to wait 831 - > on this becoming a priority? 832 - > And that's just every business and every industry is kinda 833 - > playing a gamble, right?

834 - > Of, um, I d- I, I'm not really, I don't subscribe to this way of 835 - > th- thinking in a, in a fear-based sense. 836 - > But if I just really look at, like, the math of- Yeah you can 837 - > have companies that run on this stuff, and you have companies 838 - > that don't, it's really hard to imagine how the companies that 839 - > don't will continue to stay competitive, right? 840 - > And so, um, I think one thing that's hard is, like, you don't 841 - > have the time right now, and this is gonna take time to learn 842 - > how to do- Yes to set up.

843 - > And so, uh, hopefully, you know, you have that on your priority 844 - > list, and then, or it's, it's gotta be on there soon, in my 845 - > opinion. 846 - > Audrey Chia: Yeah. 847 - > I, I think it's also, it's not a matter of, like, no time, right? 848 - > It's about whether you prioritize it.

849 - > Um, and a lot of companies, if you feel like you are okay where 850 - > you're at, you, you may be okay for now, but what happens next 851 - > when everybody is already moving onto the next wave? 852 - > Rachel Woods: The bar is moving. 853 - > Audrey Chia: Are you prepared? 854 - > Yes.

855 - > Rachel Woods: Yeah. 856 - > The bar, it's a, it's a continuously moving bar, right? 857 - > So even if you were, the, I talked, we actually talked to, 858 - > um, teams like this pretty often that were like, they felt really 859 - > good about their AI usage because they were using custom 860 - > GPTs a couple years ago, but then they've stayed at that 861 - > state, you know? 862 - > Um, and the bar has moved well, well past that, so.

863 - > Audrey Chia: Yes. 864 - > Rachel Woods: Yeah. 865 - > Audrey Chia: And what advice would you give to, let's say, a, 866 - > a founder first, so someone running their own company. 867 - > What is one piece of advice you would give them?

868 - > Rachel Woods: Um, around AI adoption or prioritizing it, or- 869 - > Audrey Chia: Yes, around AI adoption. 870 - > Like, what is one piece of advice you would tell them, like 871 - > you wish all founders would, would know or think about? 872 - > Rachel Woods: Yeah. 873 - > I mean, I would tell them so many things.

874 - > Um, yeah, it's... 875 - > I, I think the... 876 - > You know, we're lucky to work with... 877 - > You know, our clients really trust us.

878 - > Yeah. 879 - > And so they... 880 - > I know it's really weird just going back to say that the first 881 - > thing you work on, especially if you're adopting playbooking, 882 - > like your first playbook should not be your highest profile, 883 - > highest ROI playbook. 884 - > You should really prioritize momentum, and our clients trust 885 - > us and, and do that, and they are always like,"Wow, I'm so 886 - > glad we did that because now we understand.

887 - > Now we can actually make s- smart investments." Versus, you 888 - > know, I just see people really, really get stuck on a really- 889 - > Yeah hard first project, you know? 890 - > It's like, um, it's like saying, "Well, I haven't really been 891 - > hiking before, but like, let's go do, uh, that super, you know, 892 - > uh, tall one over there. 893 - > That looks like that's gonna be promising."

It's like, well, or 894 - > we could go for like a nice, you know, mild hike. 895 - > Audrey Chia: Yeah, 896 - > Rachel Woods: yeah. 897 - > You know? 898 - > Um, and so yeah, I think you'll set yourself up for success 899 - > better if you prioritize momentum.

900 - > Audrey Chia: Mm. 901 - > And what would you say to an employee who is in a company 902 - > that is navigating AI? 903 - > Rachel Woods: So also controversial, um, I think it's 904 - > a really tough place to be right now if you are waiting for your 905 - > employer to inv- invest or decide to do this stuff, because 906 - > you're choosing to move at the pace of your employer. 907 - > Yeah.

908 - > Um, this is a lot more of a personal skill set at the end of 909 - > the day. 910 - > And, you know, I, I know there's a lot of amazing companies out 911 - > there, but at the end of the day, you know, we see companies 912 - > are companies, and like you said, the, of the news around a 913 - > lot of layoffs and stuff. 914 - > And so I think, um, I just really encourage people, again, 915 - > I know it's hard, but you have to prioritize finding the time 916 - > to learn and experiment, um, and build the skill set.

917 - > And I, I just really caution people of, like, waiting and 918 - > moving at the pace that their employer chooses. 919 - > I think that's, uh, yeah, not the best approach. 920 - > Yes. 921 - > Audrey Chia: I think one thing I would add to that is, um, I, I 922 - > was previously also an employee before I moved into my, my, you 923 - > know, working on my own consultancy, right?

924 - > And the biggest change that I made personally was learning how 925 - > to learn fast. 926 - > Um, in the past, as an employee, you just wait for the brief to 927 - > come and then you work on it. 928 - > Uh, and maybe you have mentors around, maybe you are growing, 929 - > you know, together with your team, which is great. 930 - > But one thing I've also learned coming out of it is there, there 931 - > is so much to learn as a business owner, and you have to 932 - > learn the skill of learning.

933 - > Um, and I feel like a lot of employees, they may be in a 934 - > comfortable place like I was, where I felt I didn't feel a 935 - > need to learn as much. 936 - > Um, but with AI, I, I really and I strongly feel that you need to 937 - > want to start learning and feel that desire to learn so that, 938 - > like what Rachel said, you can either value add or you can be 939 - > that operator, right? 940 - > Um, or you harness the skill sets for yourself, um, and open 941 - > up new opportunities.

942 - > So AI can be both very scary but also full of beauty if you know 943 - > how to leverage it. 944 - > Um, and I think in today's podcast, we covered so many 945 - > great use cases. 946 - > So Rachel, thank you so much again for joining us. 947 - > Where can our listeners find you, and who should reach out to 948 - > you?

949 - > Rachel Woods: Yeah. 950 - > This was so fun. 951 - > Um, I always love every time that we get to jam and talk 952 - > about this stuff. 953 - > So I hope this was helpful, uh, for anybody listening, and I 954 - > would love if you want to connect with me on LinkedIn, I 955 - > hang out there all the time.

956 - > Um, our company is Amp. 957 - > Uh, our website is Run On Amp if you're interested in learning 958 - > more about playbooking. 959 - > But, uh, shoot me a message on LinkedIn. 960 - > I would love to connect.

961 - > Audrey Chia: So if you are still thinking about it, no more 962 - > sitting on the fence. 963 - > It's time to take action today. 964 - > Reach out to Rachel, learn more about what her company and her 965 - > amazing team is doing. 966 - > Thank you so much for sharing your insights, Rachel, and thank 967 - > you folks for tuning in.

968 - > Don't forget to hit the bell for more actionable AI and marketing 969 - > insights. 970 - > We'll see you next week. 971 - > Take care.