Automate Your Agency

Before AI automation, start with manual workflows

• Alane Boyd & Micah Johnson

Share episode

Send us a text

Alane and Micah tackle one of the most crucial yet overlooked principles in business automation: never automate a process you haven't done manually first. Through real examples from their own business, they demonstrate why the "manual first" approach saves time, money, and prevents costly rebuilds.

They share a detailed case study from Alane's podcast guest booking process, revealing how manual work uncovered hidden business value that would have been missed in premature automation. They explain how doing processes manually first reveals critical nuances, decision points, and the true ROI of automation projects.

This episode is essential listening for business owners who want to automate smart, not just fast.

Disclosure: Some of the links above are affiliate links. This means that at no additional cost to you, we may earn a commission if you click through and make a purchase. Thank you for supporting the podcast!

For more information, visit our website at biggestgoal.ai.

Want more valuable content and helpful tips?

Explore our Courses/Programs:

Enjoy our Free Tools:

Connect with Us:

0:00:00 - (Alane): Welcome to Automate Your Agency. Every week we bring you expert insights, practical tips, and success stories that will help you streamline your business operations and boost your growth. Let's get started on your journey to more efficient and scalable operations. All right, Micah, let's chat about something that some people don't want to talk about.

0:00:24 - (Micah): How do I know if I want to talk about it?

0:00:26 - (Alane): Well, you do. It is how important when you want to automate something. AI or not AI when you want to automate, how important it is to have a manual process that you've been doing.

0:00:38 - (Micah): I do want to talk about this.

0:00:40 - (Alane): I thought so.

0:00:41 - (Micah): Yeah, yeah. We see this all the time.

0:00:44 - (Alane): Exactly. Yes.

0:00:46 - (Micah): Yeah. So we've fallen into this trap ourselves with our last business. Even early stages in this business, before we even try to do anything, we're like, let's automate it. We're going to automate this piece. We don't even need to figure anything else out. Let's just automate it.

0:01:07 - (Alane): It's like the idea of building too far without having sold it.

0:01:12 - (Micah): Yes. I've never done that before to your team. Yeah.

0:01:17 - (Alane): But it does. It really complicates the build of an automation when you don't have a working process because there's so many unknowns or things that you realize you don't want and want to change.

0:01:30 - (Micah): Yeah. I feel like the magic word here is nuance. If you don't actually test your workflow manually, if you don't work out those issues, you don't understand all the nuances that are in place to make it work. And I don't know about you, but for me, when I think about things, thinking about them is way easier than trying to sit down, put it into a diagram, draw it out on a whiteboard, or, heaven forbid, actually implement a process.

0:02:04 - (Alane): For sure. We started doing something for our own internal automation projects and we. We do it with clients now, and it's asking questions around what the value of this is. And it's the most painful part because most people don't want to do that. And I remember even with our internal team the first time, like, the team's like, no, we could skip that. I'm like, no, we are going through this. We are going to figure it out.

0:02:30 - (Alane): And you don't know the value of a workflow that you've never done before. How do you know how much time it's going to save? How do you know the impact that it's going to result in?

0:02:39 - (Micah): You can't even make a business case around it.

0:02:41 - (Alane): Yes. So if you can't make a business case for it, then you should try doing it manual first so that you can see all the things that it's going to be valuable. Or maybe you figure out this isn't really worth doing.

0:02:54 - (Micah): Or. Yeah, I mean, we've, we've seen that many times where you start going down a path, even internally for us, and we're like, whoa, this is way more complicated than I thought. That nuance gets in the way. There's all these decisions that need to be made. We're not ready to standardize this yet, whatever it might be. We, we could look at this and go, actually, it's a little bit easier just doing this manually. For now.

0:03:22 - (Micah): Maybe we automate one piece of it, or maybe we streamline one piece of it. But geez, when we started down that path, we were thinking, let's automate this whole thing. And by the end we're like, actually, if we could just automate this one piece. That's way better.

0:03:38 - (Alane): Yeah, that's true. One, one comes to mind. You weren't in this meeting with my team, but we were. So I do a lot of guest on podcasts where I'm a guest on somebody else's podcast and every podcast is a little bit different, what their topic is, what they cover, and then where do I fit into the interview? So the podcast host might say, can you send me some bullets for our topic? And then I'll frame some questions or whatever around that.

0:04:05 - (Alane): Well, there's a, there is a nuance to that, right, Micah? Of course there's like certain things that are under my umbrella. I'm not going to talking about dental work, so it's gotta be in the realm of what I do. But you wanna make it catered to what their podcast is about. And we've been doing this manually, my, my podcast manager and I. And at the core of it. Is it a lot of time? It's not. So when we were going through making this business case for it with our team saying, I'd like to have an AI agent have access to all the podcast interviews that I've done, all the topics that I've covered that I've sent them so that the AI agent has access to these topics, and then I just put into a chat interface the new description of the new podcast that I'm going to be on, and it'll generate podcast topics for me from what I've done before and what this podcast is about.

0:04:57 - (Alane): Well, at the top level, at first, we're going through making the business case for it. And I'm like, it's not that much time. It's probably 20 minutes of time. And I'm on about a podcast a week that doesn't feel like a lot of time. But where the business case, where I was going through our questions, is that the time for me to get back to the podcast host? That might be two weeks if I need to think about it or if I need to make time for it.

0:05:25 - (Alane): Well, being on podcasts, we get leads from that. So that's. That is me waiting two weeks to get back to a podcast host is delaying leads that could be coming in.

0:05:36 - (Micah): Yeah, it's part of your sales cycle.

0:05:38 - (Alane): That's part of my sales cycle. So that at first I was like, oh man, is this even worth doing? And when we're working through it, I'm like, wait a minute. Yeah, because that's two weeks delay in a sales cycle. And sometimes it might even be longer. You know, I'm, I'm averaging it out. But that is a business case. Then it is valuable for me to have that AI agent that either my teammate that helps get me booked uses or myself, and we can get back immediately.

0:06:05 - (Micah): And I love, I feel like there's so much to unpack there, Alane, because what you. There's. What really stood out to me as you were explaining that was, Was this whole concept of if we were to take that at face value, one, you're already doing it manually. So you can start framing this and you have the basis and the understanding and the foundation to work through these decisions. But if you were to take it at face value, go 20 minutes, who cares?

0:06:34 - (Micah): 20 minutes one time a week or whatever it is. But it's probably not one time a week because you have multiple times. You're doing this even though you're just on a podcast maybe once a week, maybe more. But to get that once a week, you probably have to do it a whole bunch more. But face value, 20 minutes. Okay, should we even automate this? But then when you look at that timeline of. Wait a second, that's 20 minutes. But the delay, two weeks.

0:07:03 - (Micah): That's so cool. That's so cool. Because then you go, holy crap, if we had an AI agent doing it, then not only are we saving the 20 minutes, but that two week timeframe gets narrowed down to what, one day?

0:07:20 - (Alane): Yeah, I mean, we can get back almost immediately as soon as we are working that day. And it doesn't have to be me, which is my goal in life is not having it to be me. And so we can take the topics I've already done. We can have it. The AI agent reference it. The, the other piece that I was thinking through too, that I got reminded of in this is the credibility as a guest.

0:07:44 - (Micah): Yeah. Can I stop you right there though, Alane? Because like you're, you're identifying. We could clip out the whole beginning of this entire episode and just focus on this story. Because what you just identified was one, you did it manually. Two, you didn't take it at face value that it's just 20 minutes. Two, three, you're looking at it from what's the actual sales cycle time difference by automating this and now four, I don't have to do it myself.

0:08:14 - (Micah): Which are all angles that you can only determine if you do this manually first.

0:08:23 - (Alane): Yes, it's so true. So the last thing I was going to say is the credibility as a guest is what data I'm not looking at is. Have there been podcasts that would have booked me that didn't? Now I do keep a pretty good log and we have a follow up process, so there might not be much there. But how if, if you know, and I did not have such a tight buttoned up process for how we do things and we. If somebody says they want to schedule me, that goes into ClickUp now, that is now in my podcast interview list with an entire process there.

0:08:59 - (Alane): But if you take two weeks, does that podcast host really want to have you on their show? Now you can't even send topics. This is what your expertise. You can't send that to me immediately. And so your credibility starts to go down. So those were the things that I was thinking about at face value. Maybe not. But then when you dig in a little bit deeper and because I've been doing this manual for a year, I knew how to answer those questions as we dug in.

0:09:28 - (Micah): Yeah. Yeah. And I don't have a great way to explain this, but in my mind I see this kind of picture of when you do stuff manually for a period of time. It's almost like the cream rises to the top with the issues, the gaps. Right. So for you, it's the 20 minutes. Okay, cool. But it's also the two weeks. And it's also, I'd rather have a scalable system where I don't have to do this and I don't have to do heavy training for my team to learn how to do this. Because they can literally just chat with an agent and get what they need.

0:10:08 - (Alane): Yeah. And because we've been doing it manual for a year. If we would have built this agent a year ago, last May, when I. Last May was when I really started going after being on podcasts, I didn't have enough podcasts to fill in an agent with. I would have had three. So it would, we still would have been able to do. Would have just been a lot more AI generated and it would have been more creative on my topics and related to their description. But now we have an entire 50 podcast interviews with description of their podcast with the topics that I discuss that an agent has access to.

0:10:48 - (Micah): So let me ask you this. Cause I think this is really interesting. If you, do you think if you would've built it from the beginning without doing it manually for a period of time, that you would've ended up with the same result or the same design? Or would it have been this more, I don't know, complicated? Like, would it, would it have missed the mark and then you would have had to go back and iterate and fix it as.

0:11:14 - (Micah): As time went on?

0:11:15 - (Alane): Well, with anything with AI, if you're depending on it without the context, it's going to miss the mark and you have to. So it might have gotten me 80% of the way there and I needed to do 20% of the work, but now it nails it 99% of the time. So minutes count.

0:11:32 - (Micah): Yeah. Yeah. Well, I guess we should point out too that this is not just an agent by itself. So you bring it up. You know, this is a side tangent line, but for those that are listening, we're using a combination of technologies to power this agent, and one of them is you're saving all this data and it's being saved into a rag database or an AI friendly database, so that the agent that we designed around this can pull from that information and have that. So like what you were just saying, Alane, if we didn't have that information or if we would have just asked AI directly, we're getting generic answers.

0:12:10 - (Micah): Some of it could be hallucinations, some of it could just be made up from its training data. We don't know what it is. But because you have that information and because we took that information and provided that as a tool to the agent, now the agent can pull from that. And so you have a year's worth of information that the agent can pull from an answer to get you to that 99% not generic AI slop as the kids like to call it these days.

0:12:39 - (Alane): Yeah. And part of our ClickUp process is documenting the new podcast I was on so host description topics. Like, we have that as a step and we save that. And it was in a spreadsheet. We've changed it now that we're having an agent access it in an AI friendly database like you said. So. But we already had that because that was part of our process to document it. Because the, one of the biggest things is if somebody asks, well, what other podcasts have you been on? You don't want to have to go searching your name and coming up with which ones were the good ones? Like, we already have that all documented.

0:13:18 - (Alane): What's your topic? You're a fractional CFO and you've got a podcast about CFO stuff. Great. I've done three and I know exactly where they are.

0:13:27 - (Micah): Yeah. So you built the system first and then you applied the automation you executed on the system to make sure you knew that it worked for the use case that you were looking for. Then you applied the automation and the technology to get there.

0:13:42 - (Alane): Yeah. Your other question, Micah, about if I would have done this automation a year ago, would it have gotten me similar results? And one thing that really speaks to doing it manual first in any concept is what we do changes, what I say changes. And over the past year I have evolved on topics that I talk about. And so if I would have built that a year ago, it would have been very pigeonholed into one thing that I was doing and it wouldn't evolved with me.

0:14:15 - (Alane): Now that I feed it each time I'm on a podcast and we update the database with that information, the AI sees where I've evolved and will apply the topics to that I talk about to the new topic. So it wouldn't. If I would have done it in the beginning, it would have been so heavy on project management that it would have only spit. I would have had to mess with the prompts within the agent to give me a different result. And that's the whole point, right. Why we don't want to automate too early because you don't know what you don't know yet.

0:14:50 - (Micah): Yeah, yeah. I mean, I couldn't agree more. I see that with clients, I see that with ourselves, I see that in our past businesses is there's such a desire to go, man, I just want to automate this. I don't want to do that work that sounds tedious, but it's in that tedium that you work out the nuance. And once you work out the nuance, then you immediately can see like your vision of the. To put it a different way, your vision of, of what you wanted to automate and how you wanted it to work is probably different.

0:15:25 - (Micah): 12 months of doing it manually versus 0 months of doing it manually. That transition, you could very much articulate and go, okay, I'm saving the 20 minutes, I'm saving the two weeks, and I'm making this so I don't have to do this myself. And here's how it needs to work. Which means you could communicate that to our team and our team could go, got it. I know how to do this. And you're. You basically created a very complete requirement set.

0:15:55 - (Micah): It wasn't, I think I need this and maybe we can do this. And, you know, I'd like it to do this. Is that possible? It was like, no, I need it to go from here to here to here and do these things and fill this out and use this data that we've already created and make that into an AI friendly database. Now, you have the benefit of knowing all this stuff from doing what we're doing. Right, but the point is you're able to articulate that after doing it.

0:16:25 - (Alane): I was just going to say, like, one thing that we really talk a lot with, with projects that we're doing is can you articulate the process? Which means, have you been doing this manually? That you can really speak to it? And so, Micah, let's talk about just briefly, what it looks like when you don't and why we are really heavy on this part of it. And not everything needs to be 12 months. That was just when I felt like we had enough data and we had a good process that we.

0:16:51 - (Micah): That's when you finally got sick of doing it.

0:16:53 - (Alane): Yeah, yeah, that was when the pain point was hitting me. So not every process needs to go through 12 months of manual processes to, to.

0:17:02 - (Micah): No, it could be four weeks.

0:17:04 - (Alane): Yeah, it could be four weeks. Do it a few times.

0:17:06 - (Micah): Just get it working.

0:17:07 - (Alane): Get it working. Make sure you like it.

0:17:09 - (Micah): Yes. And so to your, to your point of what you're just saying, Alane, like this is. This has been pinging in the back of my head as we've been going down this route, but it is the fact that the process changes if you don't do it manually first. And it doesn't matter if it's two weeks or 12 months. It does matter to some degree. Where it matters is you feel confident as the process designer, the leader, the decision maker, that this process works. I can stick with this for a period of time.

0:17:42 - (Micah): You know, if that's six months or 12 months or three years, who knows? Technology is changing. But what, what you should not do at all is go, oh, I've got this idea for a process. We've never done it but I know it's going to work. Then you start building the automation or you pay somebody to build the autom and then you go, eh, that's not really how I want it, I'm going to change it. Well, you get stuck in this loop and we've talked about this in other episodes, but when your process changes, you are going down a painful journey.

0:18:18 - (Alane): Yeah. Time delays in getting the automation finished and usable for your team expense. If you're working with another company to build it for you. Because that didn't. All that development happened and needed to be paid for because you wanted to change it doesn't mean that it's free. It need, the development still needs to happen. And so what I see happen on a process that is new, you know, we're not here to argue with clients. If they're adamant about it and that's what they want to do, we will support them through that.

0:18:50 - (Alane): But what happens. And it happens every time you get to the point where you're ready to deploy the project, I mean go live with it, they have their team using it and then the team goes, this doesn't work how it needs to work. That's because they've never done it manually before. They didn't know that the what they thought was going to work. So you're at the point you're finishing a project and you realize it doesn't work.

0:19:15 - (Alane): It wasn't that it was built wrong, it was built exactly how you wanted it built. And now you got to reverse track, which in a recent project added another five weeks to the project. They couldn't figure out how they actually wanted it to be deployed. That's a time consuming piece to something that you've been wanting to launch.

0:19:35 - (Micah): Yeah, absolutely, absolutely. You know, and I think another part is just if you have a team and you're going through these rapid changes that creates burnout. So it's one thing to. And this gets into change management strategies, but it's one thing to say, hey, we're going to implement a process team. I'd like your help to figure this out. We're going to start manually, we're going to start working this out and figure out the nuance. We're going to get it working first for everybody, then we're going to start automating it. Then we're going to start adding technology like RAG and AI friendly databases and AI agents and all of that, and some of that can all be blended. But it's one thing to get team buy in to go, okay, we're, we're going to be part of this. We've got some ownership.

0:20:23 - (Micah): We see where leadership is taking this. We got it. It's another thing for leadership to go top down and go, we've now we're going to automate this and we're going to add agents. And then the team goes, okay, sounds interesting. Not sure how that's going to work, but hey, you know, you sign our checks, so let's do it. And then for leadership to go, actually, now we're going to do it this way. And then a week later, actually now we're going to do it this way.

0:20:48 - (Micah): Meanwhile, the team is struggling and scrambling just to get their work done. While all these changes are taking place, there's going to be frustration, there's going to be tension, there's going to be headaches, there's going to be mistakes, there's going to be burnout.

0:21:02 - (Alane): All right, so to wrap everything up. So what you're really looking for is what your current process is and then from that we build your future process or you.

0:21:11 - (Micah): I know, I feel like I escalated the tension.

0:21:14 - (Alane): Yeah.

0:21:15 - (Micah): So you're looking intensity.

0:21:16 - (Alane): Yeah. So what's your question?

0:21:18 - (Micah): Bring it back down.

0:21:18 - (Alane): Alane, can you document it? And from there is where you build or we build your future state where automation, AI agents can be incorporated. And I always get asked like, well, what can I do to prepare for this? And we have a free course. It's our process mapping course. And if you are trying to get the pieces together yourself or right now, you go and use that course. I think it's only 30 minutes. It's a mini course to help you understand how to process map and you can start documenting your process.

0:21:48 - (Micah): That is hands down the best way to do it. My call out for that. And this is covered in the course. But when you have a decision map both directions. That's the biggest missing piece that I see everybody make is they go if yes and then continue on. And they never map out, well, what happens if no. It's my pro tip at the end of this episode.

0:22:10 - (Alane): Thanks for listening to this episode of Automate Your Agency. We hope you're inspired to take your business to the next level. Don't forget to subscribe on your favorite podcast platform and leave us a review. Your feedback helps us improve and reach more listeners. If you're looking for more resources, visit our website at biggestgoal.ai for free content and tools for automating your business. Join us next week as we dive into more ways to automate and scale your business.

0:22:35 - (Alane): Bye for now.

People on this episode

Head Shot

Alane Boyd

Co-host
Head Shot

Micah Johnson

Co-host