Automation for Anyone: Simple AI Workflows That Save You Time

In this episode of AI for South Africa, Maxine breaks down automation in a way that feels simple, practical, and possible — even if you are not technical.

Automation is not only for big companies, developers, or IT departments. It can start with something as simple as saving your AI context, reusing your best prompts, or connecting the apps you already use.

Maxine walks through a simple three-step “automation ladder”:

1. Set up your AI tools once with the right context using Claude Projects, Custom GPTs, or Gemini Gems.
2. Build a reusable prompt library so you stop rewriting the same prompts from scratch.
3. Connect your apps using tools like Zapier, Make, or n8n so repeated tasks can happen automatically.

You’ll also hear practical examples for a small accounting business, a teacher, and a freelance designer, showing how automation can save time, reduce blank-page frustration, and make everyday work feel lighter.

This episode is for business owners, teachers, freelancers, employees, and anyone in South Africa who wants to start using AI and automation in a simple, useful way.

Transcript

0:07

Max. This is AI for South Africa. Hello, hello, and welcome to AI for South Africa. I am Maxine, your host of this podcast. So I need to tell you something a little embarrassing. Last week, I'm not exaggerating, I typed the same email four times. Four times. Different clients each time. The same structure, the same tone, the same information, the same sign-off. And somewhere between email three and four, I stopped mid-sentence to be like, hang on a second. And I stared at the screen and I was like, I've written this exact same thing before. I knew the words, I knew the paragraphs, I knew everything about it. I was essentially a human copy-paste function at that point. And I thought, why am I still doing this manually? And the honest answer is because I hadn't stopped to ask myself the question. This is really what automation is. It's not a techie thing, it's not something big companies and IT departments get to use. It's just stopping and asking yourself, why am I doing this the same way from scratch every single time? Can I automate it? And today, that's what we're going to unpack together. Recalling this episode automation for anyone. And I mean that. Whether you're a business owner, a teacher, a freelancer, or someone who's just starting out with AI tools for the first time, there's something in today's episode that you can use this week. This week. Not months after you've done a course or read up five million articles this week. So we're going to climb what I call the automation ladder. Three steps. Starting from the simplest one you can do today, all the way through to apps talking to each other automatically while you're doing something entirely different, making coffee or making tea. We're going to see exactly what it looks like for three different people. Um, because the context of your life, your work, actually matters when it comes to AI and automation. The automation that helps a teacher looks different to what it looks like for a freelance designer. And that's fine. It's supposed to, AI and automation are supposed to work differently for each person. Okay, so I want to start off by shifting something in the way we think about automation. Because when most people hear the word, they picture like robots and um like a production line or code or something complicated that a developer has to build for you and cost a fortune. Now, yes, there are those versions that exist and for reasons, but that's on step 47 of the ladder. And we're not there today. We're starting with step one. And step one is so simple you might actually be annoyed when you hear it. Step one, set it up once and save it forever. Guess what I've noticed with almost everyone I talk to who use AI tools and who have done a basic prompting course, whether it's ChatGPT, Claude, or Gemini, they open a fresh conversation every single time. And the first thing they do is explain who they are and then what their business does and what tone they want and what format they want. And and that's good because that's a good structure for a prompt. But it's like calling a colleague every morning and reintroducing yourself every single time before you ask a question. So step one is about stopping that. It's about giving your AI tools context once and then having them remember it every single time you come back. Claude has a feature called projects, ChatGPT has custom GPTs, Gemini has gems. They all do the same functional thing. They let you set up persistent context that AI carries into every conversation. So instead of starting a chat every single time, repeating yourself, giving it context of what you do where you're from, you can put it into a project gem or GPT, and from that point forward, you will have memory or context from the conversation you're starting. Now, obviously, this doesn't have to be just for your whole business or the bigger picture, it can be for small projects as well. Every project can have a different context. So let's say you want context on a building project you have, or you want to give it context on your branding for your business. Um, let's say you want to give it context on a specific subject you're writing, all of those things can have context within these little elements, so projects, GPTs, and gems, and then every time you don't have to start from scratch. This is automation of you specifically automating yourself out of repeating your own background information over and over. Now, obviously, when you have a paid-for account, most of these chat tools do remember some of the elements that you are, but it'll be all over the show. So it's very, very nice if you have multiple projects and multiple subjects or elements that you want to give it context to, and then it keeps it into that little bubble. And this is genuinely the highest return, lowest effort thing that people can do today. It takes like 10 minutes, you set it up and you save it, and every time after that you want to create a chat, you go to that project and you start a chat. Now, I want to be honest with you here, because that's kind of the whole spirit of this podcast. When I first set up my Clawed project properly, I sat there going, Why didn't I do this six months ago? Like, why did I wait so long? I was a little bit annoyed with myself having to go over and over and explain the context and starting from scratch. So don't be like me. Jump in, create a project, custom GPT or gem, and automate that part where you keep having to repeat yourself. Step two is about treating your best prompts like assets. So most of us, when we write great prompts, we get great results, we just move on, the conversation closes, or it goes into that long list of unfiled, unfiltered chats, and we forget about it. And then the next time we need, we do the same thing and we start from scratch. It's like writing a recipe once, cooking an amazing meal, and then throwing the recipe away, and then starting from scratch the next time. So your prompts are templates, and the best way, the best ones are worth saving and reusing. Here's what reusable prompts actually looks like. Let's say you send a payment reminder email to a client every month. Instead of typing the email every single time and editing the results, you build a proper prompt once with placeholders and variables in it. You save it somewhere, then you can find it. Something like this. Write a professional friendly payment reminder email for a client named, and then you put a little block where it says client name. The outstanding amount is, and you put a little block with them, the amount. The original due date has then the actual word date that you know that you need to put the dates in there. The tone should be firm but not aggressive, keeping it under 150 words and sign off with your name and your business name. And to do that and save it somewhere is so much easier than typing the whole thing out every single time you want to create a prompt. Especially when you're looking at Excel's and things like that. When you're getting into a process of getting out information or writing a report from an Excel or creating a PowerPoint presentation or creating a lecture or a theme for school, whatever it is, when you get good results, at the end of the chat, say to it, please give me a prompt template that I can save. And you save it somewhere, it can be in a Word document, in an Excel, wherever it is. And then the next time you want to produce the same thing, you can go to that document and just copy and paste, and you know that you're going to get good results. There's no point in having the same conversation over and over and over again. Because remember, with these chats, you will get different results to a certain degree if you don't exactly stipulate how you want it. And when you create that template, that's what you're doing. That's step two. It's not a technical thing at all. It's just treating your best prompts or your best outputs the way a good lawyer treats their document templates. Okay? You don't need to rewrite the whole thing every time. Give yourself a starting point to save yourself a couple of minutes. Um, you can go make a coffee with that or listen to my next podcast. So you can save your prompt library. I have mine on Google Drive and a Google Documents. Some people use it in their notes app, some people actually use Notion, wherever it is, just save it. Don't keep rewriting your good prompts, especially if you want consistent results. Then there is step three, and this is where apps talk to apps. So you can use things like Zapia Make or even using N8N, it's a little bit more technical, but you can. And step three is where it starts to feel a little bit more like magic, but I promise it's not that complicated. So Zapia and Make are what you call automation platforms. And what they do is they connect apps to apps. So when something happens in one app, something automatically happens in another app without you having to lift a finger or do anything. Think about an app that you already use. So like Google Forms, Gmail, Google Sheets, WhatsApp business, maybe you use Slack. These apps don't naturally talk to each other. But Zapia sits in the middle and creates a bridge between the two without any technical knowledge or need. It's literally a drop and drag. So the way that it works is you'll get a trigger and then there's an action. So the trigger could be the first thing that happens, and then the action automatically happens as a result of that trigger. So, Max, what exactly do you mean? You're losing us here. So let's say a client fills in a Google form, right? Zapia acts as the automatic bridge and it goes add these details to a Google Sheet and send a welcome email the moment that form is filled in. All in one go automatically. I don't want to do anything. That's the kind of idea of the automation part. Zapia has a free tier, which is usually enough to test one or two simple automations before you commit to paying anything. But the moment you start using it and you you see the results, um, it's worth every penny. Then there's make, which previously used to be called Integra Mat. This is a little bit more powerful option. Um, I prefer the interface, it's a lot more easier to use than Zapia, and it's a little bit more flexible, it's slightly steeper learning curve in terms of um there's so many templates that you can use. I sometimes feel I'm a little bit overwhelmed by the templates, but once you figure out how the system works, then it's really, really easy to use. Then you have N8N, which is a little bit more technical, um, and there's a lot more you can do with it in terms of connecting it to Claude and Um Chat GPT. I won't get into too much information, but if you are technically oriented and you want to dive into a tool that's really robust, N8N is a good option. Now, I want to be real with you, right? Step three does have more setup time than step one or two, and you'll probably need to watch like one or two tutorials to see how the program works, depending on which one you use. Um you got to be able to figure it out a little bit, trial and error in step three, but it really does make a difference if you're looking at something to automate, it's worth it once it's set up. So, quick summary on the three steps, right? Just so we can get the whole picture. Step one, you set up your AI tools with context on a subject or on a topic. You can use clawed projects, custom GPTs and gems, and you can do that in 10 minutes. Step two is build a prompt library, save your best prompts with placeholder variables. And what I mean by placeholder variables is like you can put it in brackets and then you actually put what you need to put in there. So if you say the date needs to be in there, you add brackets and you write the word date so that next time when you go there, you'll remember that you actually need to put the dates in there. So that's what placeholder variables means. Then once you have those prompts, we're reusing them instead of starting from scratch, and you can start using that today as well. The moment you create a good prompt, save it tomorrow. You can use it again. The third one is connecting your apps with Zapia or Make or N8N, trigger an action, and then everything happens from there. It follows automatically. I will be sharing a couple of articles as well on my LinkedIn newsletter with regards to how to set up Zapia and make. You can follow me on LinkedIn or on Instagram to find those deep dive tutorials. And if you want help from me, I'm more than happy to help you out. We can set up a session and we can see what the business need is and take it from there. Most of you listening right now are super okay to jump into step one and step two, and it's generally the best place to start. You'll already be freeing a lot of time. Right, so let's look at some examples of how these things actually work and how everything sits together. So there's a lady, Fatima, a small accounting service in Cape Town, and she was using step one and step two. She was running a small accounting practice. She had all about 18 small businesses, sole traders, small retailers, a couple of professionals here and there. But she had good steady business. But every month she does something that quietly steals her time. She personally writes, your managed accounts are ready to every single client. Different name, different figures, but the structure is always identical, and the tone is always the same, and the sign-off is always the same, and the estimates. She spends about 45 minutes on those emails at the end of the month. Probably a little bit more, give or take. What she did was she set up a Claude project, gave her firms context so that it understood what she wanted to achieve. She gave her business name, what kind of client she works with, how she would like it to be heard, a professional but approachable tone, and a note that she prefers short paragraphs and no jargon. Okay, so she was doing step one, boom, done. Then she built herself a reusable prompt that she could just copy-paste every month and get the same results. And that was step two. And at the end of the month, now Fatima opens up her Claude project and it knows what she's trying to achieve. She has a prompt template that she fills in the details for, maybe takes her 20 seconds. But instead of taking 45 minutes to an hour, it takes her eight minutes to run through all the appliance and it's done. Boom. So that's like a perfect example of where you should use these things. Another example is there was a teacher, Mrs. Glamini, and she had 38 learners in her class, and every term she would write individual progress comments for every single one of them. You know, the ones that go at the end of the report with a little motivational thing. I mean, obviously, she knows every child, she wants there to be some kind of personal touch to it. So what she did was she built herself a prompt and kept in place certain elements, but then also gave her space to give a little personal comment at the end of it. The bulk of it, the 80%, was done by the prompt, the learner's name, what their strengths were, where they need encouragement, and then her personal note on what the child did for that term. So that was still personalized, but the bulk of the work was done. She saved the prompt, and now every time reporting comes around, she uses the prompt, she fills in the name, the strengths, and the development areas, she adds her little final notes, and she saves herself so much time. She more had a problem of like this blank screen situation, you know. So even though she was saving time, she just for 38 people, you've got to come up with creative stuff to write. So she thought, okay, well, we'll just keep it as a prompt, keep it as a template, and yeah, she does it all in 45 minutes, which is fantastic. And to me, that's what automation is supposed to do. It must never do the 100%, it must take you to the 80, and you give it your chutzpah of 20%, you know. Right, so then the last one I have is a graphic designer, Kianu, and he also had a good client base and he was juggling a lot of things. And every time I came to invoicing, he was manually sending WhatsApps to his clients every time to remind them that they needed to pay. He decided he was going to create a Google Sheet to track the project and see who had paid. So he created some columns to track it, and then he connected Zapia to Google Sheets, and every time an invoice status changed, it would send the client to WhatsApp automatically. So Zapia automatically sends the WhatsApp message to the client based on the status. He also created a template, a prompt template that it can use every time, and in no time it was already set up and to go. He didn't have to write the messages every time, he didn't have to touch anything. So the spreadsheet is there, the WhatsApp and flow is there, and now he's incredibly organized and he doesn't have to worry about that task anymore. Again, another example of if you can automate it, automate it. So what is your action for this week? Right? I'm genuinely asking just to take one thing, not all three. I mean, you might not be at step three. You might be okay. I'm not even at step two, Max. Maybe you're only at step one, and that's fine. But try it, see what comes out. If you don't like it, that's okay. But this week I want you to find the most repeatable task, the thing you do the same way from scratch every week, the email you write, or the chat that you have with Claude about the same project. Write it down. Remember, we like to write things down so that we can understand exactly what we want. And then ask yourself is it a step one, step two, or step three problem? Then start looking at the tools that you want to use. Decide if you want to use Claude Projects or a GPT, if you're already using prompts regularly and you're producing stuff, maybe you can go back and get those prompts into your library now. Then it's a step two. If you're more comfortable with diving in and trying new technology and not being afraid that things break, try step three. There's no such thing as failing. We're trying and we're learning, and we're in this together. So just pick one, take action, share it on the social medias, and let's see what we can produce. If you get stuck as well, drop it in the social medias and I will help you where I can. This is gonna take you like 20 minutes to try. Give it a bash. So that's a wrap on episode four. We walked up the automation ladder together and setting up your AI context once and never repeating yourself or building a prompt library or creating zaps with Zap here. And we actually saw with the three examples how simple it can be and how much it can save you, not only from a time perspective, but also from a blank page perspective and a frustration perspective. So none of these require developers, which is great. None of them require budget if you're starting out. So yeah. I hope you're sitting and asking, why am I still doing this the hard way? Can I help myself with this process? The next episode, we're going to be stepping back and asking, before you bring AI into your business, what should you think about first? We'll also talk about shadow AI, data, change management, and question leaders need to know before you start buying these tools and implementing it into your business. Subscribe so that you don't miss the next episode. And if you learned something from today's episode, please share it with the person who you think is gonna need it as well. And I'm Maxine, and this is AI for South Africa. Let's go.