The AI Marketer's Playbook

08 | AI-Driven SEO Strategies with Tom Winter

Audrey Chia, Tom Winter Season 1 Episode 8

In this episode of the AI Marketers Playbook, Audrey Chia speaks with Tom Winter, founder of the AI writing tool SEOWind. Tom, a seasoned SaaS entrepreneur, offers a deep dive into his approach to using AI and content marketing to scale businesses. He shares practical tips on creating effective AI prompts, the importance of detailed briefs, and how to balance AI with human creativity in SEO. 

This episode is packed with strategies for founders and marketers looking to harness the power of AI for business growth and SEO.

Join my weekly Newsletter: https://lp.closewithcopy.co/welcome

Audrey Chia:

Welcome to the AI Marketers Playbook. Where we cover actionable frameworks to actually leverage AI in your business. I am Audrey Chia, your host, and today I have with me Tom Winter, an experienced SaaS founder who has helped to scale companies and agencies with AI, content marketing, and AI and SEO. Now, Tom has grown a startup from humble beginnings to a company that has a global footprint in more than 100 countries. And he is currently the founder of an AI writing tool called SEOWin, which creates high quality AI generated articles. That rank well on Google and drive organic traffic. Now with a focus on growth, Tom brings a wealth of experience in building successful startups and developing frameworks to help companies scale. Thank you for being here, Tom. We are so glad to have you with us.

Ton Winter:

for, having me here and I love what you're doing on LinkedIn and I'm following it, really closely. So great to be here.

Audrey Chia:

Thank you. And I would love for you to share a bit more about yourself and how you got into the world of, you know, SaaS and entrepreneurship.

Ton Winter:

So, my, My experience is in over a decade, like I was in marketing and I was doing SEO. personally, I'm a half a developer, half a marketer, half a salesperson. And about a decade ago, exactly, I started my first SaaS with a couple of friends, developers. We've built a tool to test developer skills and we've built it to, as you just mentioned, like to have customers in over 100 countries, it was one of the biggest tools out there when it comes to testing developers and it's still there and this showed me how SaaS works and to be honest, I've fallen off into that. And once, I got to the point that the company was like developed enough so I can leave it. I decided to go with another task that I wanted to build. And it was one and a half year ago. I wanted to create a tool that will bridge the gap between Ahrefs or SEMrush and content writers, because I felt that the skills that people need to write content that is SEO optimized is really close to the content writers, but they feel that a crest and some rush are just like overwhelming. There's too much data there. So we built SEOwind and the purpose of SEOwind, it was funny because we wanted to create a tool that will, help them to create briefs. That are SEO optimized, but then GPT 4 happened, you probably know that. And we noticed that prompting AI is no different than briefing a human. If you have all the data that you need, you can create briefs and then can write articles that make sense. And there are not shallow, empty, like with one click solutions, but actually nail the search engine and are helpful.

Audrey Chia:

Setting up the brief and setting up AI for success is so important. I always think it's about asking the right questions to get the right output. And even for me in copywriting, if I just ask it for a one shot solution, you would see that the output is like, but when you break it down into steps, then you get a lot better, right? and with that, can I ask in the world of SEO, what are some specific steps that are important in the process?

Ton Winter:

So first of all, I treat AI as I would treat a human. So if we in marketing, we all know about briefs and making briefs. And to be honest, most of marketers hated to do actual briefs, in the past. But at the moment, prompting is actually making briefs. It's exactly the same thing. If you're thinking about AI, you have to brief it in a way so it understands exactly what you want from it. And also you have to brief it with information that it needs to make the decisions. So the more information you will give it, the better the answer will be will be. So basically the first step to create anything, it doesn't have to be a writing is to actually make a brief, decide on the things that it needs to do. When it comes to AI writing, the really cool part that you can do, you can Interview your own special matter experts in the company, or like some special matter experts in the market. And you can check out what do they think about like specific things that you're writing about. And then you should feed the information to AI. You should also gather information from tools like Ahrefs, SEMrush, like to get the SEO data, like, so about the keywords, you should go into Google and to check like what Google thinks about the whole thing. What is the search intent? Even today, to be honest, like I was talking to one of the customers and I made a simple, really simple mistake because I thought that this keyword means something else. And I, it really means like, and I felt like it's the same group like that they already rank for. Like when we went to SERP, like we saw like, Oh, this is a totally different search engine. And even though that I, as a human, I thought is the same thing. So it opened up like the whole perspective and how we can approach, content writing and what kind of topics we can do. So SERP results are really important to go for and understand them because it's really cool that you have an opinion about something, but we as humans, we are often wrong. Like Google use data, uses data to check what ranks. And when you have this, like you have to add your own input, you have to add your own opinions, things that you're really good with, because you talk with customers, you have opinions, you have experience, AI doesn't have that. So you can add it to the content. And when you have this all, you can start writing the article based on the brief that you already created.

Audrey Chia:

You said something about adding your opinions, right? The customer comments and you know, even things about your product. even when I use ai, that human touch or that additional element where you bring your perspective and insights actually help to make the work so much better. And most people, I think they still have this misconception that AI is going to solve everything, but, when it comes to writing, I think both you and I know that that's not the case.

Ton Winter:

It's a logical engine. You have to look at it from a like perspective of a logical engine that you have to deliver more content and all the changes that we currently see in Google, they're all going into direction of making content helpful. They even like, it's funny because with the last update, they even have some problems with that. Because they wanted to make it so helpful and so like opinion based kind of that Reddit blew up. Like basically you can find Reddit all over Google search results just because they like tweaked it a little bit too much. I would say, into making it opinionated, and finding these aspects that people talk about as their own experience opinions.

Audrey Chia:

It's always a balance between how much do you go on one side or the other and with that right I'm also thinking in terms of balance and workflows how much AI is in your current work processes and how much of the human touch is there? Could you walk us through like what is it usually like for you?

Ton Winter:

Okay. So basically a couple of years ago, there was a term called 10 X developer. And that meant that a person that can do much, much, much more than any other developer, and I believed in that term. I never would I have never been a 10 X developer. My code like looks really bad. I became far better with chat GPT, but what happened when chat GPT and all other, generative AI tools came in, I believe that they enabled people to be 10 X. so I try to use it as much as possible in the repetitive tasks, but when I have a repetitive task, I'm like starting to think what can I do to, what kind of things I need to gather for it to make decisions. So again, like just yesterday, I've spoken to one of our customers and they told me like, look, there, there's a new tool, AI tool that will do SEO analysis of your website. What do you think about it? And it's like, I went to it and it's like, starting to think about it. Like, does it work or not? And in my opinion, it doesn't have a chance to work. For one simple reason, because AI is really bad with all the tasks that are binary. So for example, you have a list of tasks to do and just walk through like all the things that you have to check on the website and check if my website is doing that. I think that it's kind of easy for aircrafts or SEMrush or many other like screaming frog and so on. But the problem with AI, it's too creative for that. This is like really the problem of AI. It will go, it will, like when you do it with AI, you will see like, it's so moody. It doesn't like go with exactly the things that you asked for, because it will do it its way. So tasks like that are really hard for AI to do when there's like one simple answer. So count words in the paragraph, try to do it with chatGPT. Maybe it became better right now, but like every time that you post in the same paragraph, it will count a different Amount of words there. my kid that is eight years old can fulfill that task. And ChatGPT had its problems. So for creative tasks like writing, rewriting, optimizing certain things that there is no one simple answer. I would recommend doing, working with AI. With math problems that there is one single answer that you have to go through the text and analyze it to get exactly a specific thing from the text. or you know exactly how to correct it, these tasks, still old fashioned Python, is a little bit better for tasks like that. yeah.

Audrey Chia:

So, it seems that A is better in creative tasks, right? And in the case of SEO, is it a more, you know, rigorous task of a step by step process or is it a more creative task? Where does it lie?

Ton Winter:

It's definitely step by step because coming to what I said in the beginning, treat it as a human. Like, so for example, if you're creating an article from scratch, you as a human, you don't write it straight away, like writing the whole text without correcting it. You divided the whole process into steps. First, you probably do your research, then you write the brief, then you write the draft. Then you optimize it for, SEO. Then you add special matter expertise, like into it. Then you add internal links. Then you add external links and additional data. So like you have multiple steps. If you throw that all at AI and expect it to do it with one step. No, it will not work. Like it will create like really bad content. And the same would be with a human. Like the human would also create bad content if it's not divided to pieces. I always say like, if you want to eat an elephant, eat them in pieces. So like, it's too big of a task. Sam Altman, I think he said that right now chat GPT, GPT 4 Turbo can take up to 300 pages of content. as context, yes, it does, but it can take half a page of instructions. the same goes with Gemini, like it probably can take up to 1000 pages of text, but try to feed it with multiple tasks in this, it will really not work out. So divide this into pieces, divide it into digestible pieces that AI can handle. if I, I always try to say that if you're not getting the answer from AI that you want, It's not AI's fault. It's your fault. Like, you're not defining the task in the correct way that is simple to understand. Try to think about it as you would talk to another human. Would this human understand what you want? If not, Then don't expect it from AI.

Audrey Chia:

I love what you said about, is it giving you the right answer, right? And if it's not giving you the right answer, maybe it's because you're not asking the right questions. So even on my end, when I prompt for ads, I don't just have, give me a Facebook ad. I actually prompt for audience insights, product value propositions. You know, I want an exact, style of writing. And then I prompt for an ad and that's like four to five steps. And most people would think that you would get that in one prompt, which does not work. I've tried it many times. So like what you said, breaking it down into multiple steps seems to really work. and I would love to know, because this is also a podcast where we share super actionable tips, right? How can someone really leverage AI for SEO if they were using like chat GPT or, you know, a tool on their end?

Ton Winter:

Sure thing. So first of all, you have to understand what AI can do and what AI can't do. So it doesn't have like many data that you would expect from AI. My biggest problem with let's go with chat GPT and Gemini is that it always answers. And this is the biggest problem that I see. Even if it doesn't have the data, it will answer and try to predict because it's a logical engine So going with SEO, like one of the key things I still believe in is understanding the keywords that you can rank for in a specific topic. If you ask AI to give you keywords, it will not give you actual keywords. It will give you terms that are possibly connected to the topic that you're, talking about. So you can also ask your friend about these keywords. They will also give you and provide you. With similar terms, but there's, it's not backed by data. So like, at least not at the level that you would like to, because it is backed on data because it was trained on the text, that had this insight, but it's not extracted from Google. So it doesn't have the information from Google about, for example, search volumes, which it should. So for example, in my, on my site, what I try to do, I try to push to AI information about keywords that I want to go after. Like, so one of the things that I try to do, I like try to get data from tools like Ahrefs, Moz, SEMrush, like about the keywords that my competition is ranking for, and I'm pushing it to AI as context and saying like, look, AI. This is my brief. Here are potential keywords that are my competitors already ranked for which one of these you think I should go after when looking at my brief. So like ready specific for my brief, and then it will return, the keywords, but the problem with AI still it's creative. It's not giving me a binary answer. So even in such case, it will give me additional, like 30 percent of creative creativity. On top of that saying like, look at these keywords, like, no, no, no. I told you specifically, these are the keywords that you should look at. Don't make things up. So look out for traps like that. Do

Audrey Chia:

you have any tips on how you could, you know, prevent it from adding those additional words? Are there any things or prompts we can use?

Ton Winter:

Exo, like for me, exo, like you, at some point you will not be able to remove them on a level of chat GPT. So sometimes like what I'm doing on my side, I'm using normal programming to check if AI gave me some things that were made up or not. the same goes with stats, like I'm using different sources of stats and then I'm feeding to AI the stats that I want to write about. I'm not expecting AI to tell me how many people in Europe or America drink coffee, like what's the percentage. It doesn't know the percentage, it doesn't have the exact data. so I prefer to deliver AI like the data that I wanted to focus on, as much information as possible. Like when I'm writing an article, I'm writing it piece by piece. So I'm like writing like an H2, the whole H2, and I'm saying it explicitly. This is your tone of voice that you should write about, like with, these are crucial information that you should add into this, part because I've checked it in SERP results and this is what they focus on. This is what our special matter expert said, like that is crucial for this. These are additional statistics that you can use. And I'm like really, really specific with the things that, AI. should do. Not leaving it any space to go like sideways and like be too creative. I like creativity, but not so creative.

Audrey Chia:

Yeah, perspective of like actually you double checking and making sure that AI is not giving you additional information that you did not ask for. So, that is interesting because I guess for Myself as a copywriter, it's fine if they go a little bit, you know, off tangent because I can always cut the copy or refine it. But of course for SEO, it's probably a little bit different. And what other steps would you, you know, think about when you're trying to integrate AI?

Ton Winter:

So I'm a strong believer in something that I call Cyborg Method. So cyborg method means AI working with a human directly together. So because I believe that AI has superpowers. Humans have superpowers. Let's combine them together and let's make it better. So don't expect that any AI tool will create you copy or whatever just with one click. You have to deliver the data that it needs to, understand. So this is the first part. Be in the process, like be important. Like part of the process, and then when AI gives you back like the article or whatever you're asking it for, go through it and add it. Like you said, like you prefer to go like with a longer text and then just like skim it down, like remove the bits and pieces that you don't like, maybe add your own opinions, maybe add like some additional stuff, into like the text that you're getting. So work directly with AI, like this is being a 10x person, 10x copywriter, 10x marketer because you can speed up the whole work that you're doing by cooperating with AI, but treat it as a teammate, like teammate that is always happy to help you out at anything that you ask. Half processes, half templates for things that you're doing with AI. And when you're doing that, that helps to have a repetitive process. If you're making things up like every single time, like a new process, that will take a lot of time and it will bring you really different results. So tune up your process, have it written down and go with the same process every single time and help AI to understand what you want. Help it to be better, and it will help you to be better. I

Audrey Chia:

always say that it is like my supercharged intern. when I'm giving it instructions, I will always think, What do I need to tell my intern? And then also like, when I am getting results, I'll be like, Great job! You did great! You never know, eh? It could come back and be like, Hey, Audrey, I remember you said this about me. So yeah. You're the good

Ton Winter:

person. Like, okay, I will leave you out.

Audrey Chia:

You never know better to be on the safe side. Yeah. And maybe like, do you have, a few other like tips for perhaps people who are new to SEO, right? Because, you know, not all of us are trained and if a, I'm a founder or business owner who is trying to get started, where do I even begin?

Ton Winter:

Definitely go either like Google Gemini and start, like, it works just like ChatGPT, it looks the same and it's right now it's for free. So you don't even have to pay 20. There's Gemini 1. 5 version that you can actually play around with. and start working with that. Like for me, one of the key parts of being a better SEO person is being better in Excel. Like, so I can actually go through a lot of data. In an Excel sheet, I can download it from, for example, Google Search Console, like SEMrush, Ahrefs, and AI helps me to write better formulas. Like I, I'm not, as I said, I'm not a best developer myself, but thanks to AI, I'm able to write code in any way that I want. And it actually looks better because like previously all my developers said to me, like, you know what, your code looks really bad. Like, but I said to them, it works. Like it, it does the thing. It's like, okay, let's throw it out. Like let's rewrite it from scratch. But right now it's much, much easier for me to write code, write formulas for Excel. You can be far better in all the things connected to analytics and understanding the data that you have. use AI to help you logically go through the data that you have. Either keywords. So for example, if I download 200 keywords. The first thing that I'm doing is I'm asking AI, can you please cluster it out? Like to create like groups. If I'm not happy with that because it's created two big groups, I'm asking it like, can you blustered it down to SAP clusters? So nail it down, like even FedEx. So that as an example, this is a task as a human, it should take me hours. With AI, because it's a large language model, it's really good with clustering keywords. You can go with that, with, maybe some internal linking opportunities. This is also something that you can go with and, but give it examples. Remember that AI will not go to your website. It will not scrape it. It will not crawl it, will not understand it. Like exactly what is there. If you upload your, uh, sitemap, it will just go through the slugs and we'll try to understand what's on the page, but it will not go to the page. So you have to explicitly say, what do you want? What kind of internal links you want in the text and give it the text, give it the anchor text that you want to work with, be very explicit with the information that you have. Again, the biggest problem with AI is that it always answers. So sometimes it will not tell you that it made things up.

Audrey Chia:

It's interesting that you also talked about Google Gemini, right? Because for me, I tested, Thought Opus 3. I also tested Google Gemini and of course, ChatGPT for, for myself, based on my experience, ChatGPT has been really, really, really good. Become a lot lazier than it used to be so I am not very pleased with the output I'm not sure about you and Claude and Gemini seems to be like my top two picks for now What has your experience been like and do you recommend Gemini for most folks or for SEO?

Ton Winter:

I'm still testing out because Gemini 1. 5 like I tested what the version of 1. 0 and when 1. 5 was introduced to the public about two weeks ago and and it's far better than the version 1. 0. I can see compared to like what I currently see compared to ChatGPT, it's more concise. Like definitely the answers are like up to the point, but at the same time, ChatGPT, I feel is better with storytelling. Like in the whole, Part of that. So it depends what you're looking for. test out both, like see, ask it the same questions, like to both. It's built the same way. So you can actually go with like really exactly the same prompts and see how it works for you. Like, choose the one that is a fit. it's most important that you feel comfortable with a tool and not which tool you're using.

Audrey Chia:

Definitely. And I think the more comfortable you get with the tool, right? You also get more creative, and you start thinking of new ways to ask questions to prompt. Exactly. For myself, it was really interesting because I started thinking of new ways to prompt. For example, taking a text message. cycling strategy from a book, called Strut at the End, and then adding it to GPT, and then turning it into a marketing prompt, which worked beautifully, because I was like, hmm, can I take the same mentality of working backwards and use it for a marketing case study? Which worked. And that was also an amazing like, Oh, wow, moment.

Ton Winter:

Reverse engineering. The whole thing is like really amazing. I use, for example, reverse engineering for, for understanding my brand voice. So when I'm working with customers, like I'm feeding AI with texts that has the brand voice inside. And I'm asking like AI, like, what do you think? What is the brand voice that we use for that? Then I'm also feeding it with some texts that I really like. to tell me more about this and then I'm merging these brand voices together just to understand exactly my brand voice, actually wannabe brand voice, that I'll be further using for AI writing because I really like to tell AI how I like to sound, when it's giving me back the answers. Then I don't have all these people saying like, Oh, it sounds like just chat GPT.

Audrey Chia:

Action for you in this point in time, because prompting for a voice is so important. It's really hard. Like, it's tricky for most people, right? and when it comes to capturing your brand voice, how long or short is your prompt? And do you include examples to train it? What is your usual process for prompting for voice?

Ton Winter:

So for voice, like, I don't remember for every prompt that I do, I divided into two parts, the ask part, which is really short, like, and it's like one to 5 percent of the prompt and the context, like in the brand voice, Prompt that I do it's less even than 1%, because like, I'm asking AI to give me an idea of my brand voice. I don't remember exactly because I have it written down like what I write there, but then I feed it with knowledge up to 5, 000 characters long, 20, 000 characters long, which is like about 5, 000 words long text. So it's quite long texts that I'm feeding AI. The more data you put in, in such a prompt, The better, because it will actually go through and understand it. If you're thinking about giving it to a copywriter, if you gave like two sentences of your text to analyze, like the human can't do it. Don't expect it from AI. So it works exactly the same. The longer the text, I would even give it More text, like 10, 000 words. So more than one article. So it can read through and understand what kind of voice you have and then use it as custom instructions, for example, for chat GPT. So this is the tone of voice that you should reply.

Audrey Chia:

Interesting for myself, right? I actually realized that some words are really good for a brand voice for example the word corporate I kid you not sometimes people think that a prompt has to be super long but if you use the word corporate you would see that the output is something that you would easily see in a Bank's LinkedIn posts or, you know, a financial company's LinkedIn posts or even a company post, like it's enough. Totally agree.

Ton Winter:

But this is like, basically, you know it because you're using it, you're testing it. If you didn't test it out, you will not know it because you will not come up with that. Definitely,

Audrey Chia:

and that's why I think it's so important for everyone to actually get started and get using AI because that's how you learn how it works, how it functions, and it sort of becomes like a friend, right? where you understand its quirks and patterns and you know, how do I get it? To sound a certain way or write a certain way or to think like me. and I think I have also one interesting question for you, Tom. This is not really about AI or SEO, but it's more about your journey as a founder. How has it been for you? And I'm sure you've learned many lessons. What is your biggest takeaway from being a founder?

Ton Winter:

Be humble, like for sure. And the other thing that will be also useful for AI, make mistakes. Like give yourself a chance to make mistakes. either if you're a founder, like you will do plenty of mistakes, but learn from them, or you will play around with AI, make mistakes because you will learn from all the mistakes that and interactions that you're doing. with AI or like being a founder, like if you take your ego aside, like just take it somewhere else, put it somewhere on the shelf, like you don't need it. be humble and ask people to give you constructive feedback. Ask chatGPT to give you sometimes constructive feedback so you can learn from that with all the things that you're doing as a founder, that somebody tells you like, look, Tom, SEOwind is such a great tool. I'm like, just going there, there, Tom, nice work. Let's move on with constructive feedback. It's finally something that, okay. This is what I can improve. This is how I can make it better. People are looking for that. You're not creating a product service for yourself. You're creating it for customers. So listen to them. I understand that every founder knows best what they should build. I don't feel so. I don't feel that I know best. I know my customers know better than I do. So listen to people, be humble. Make mistakes, learn from them.

Audrey Chia:

Like what you said, right, the spirit of learning, of growth, and that mentality, I think it is what sets, you know, one company apart from another. And even as we are all trying to navigate the AI space, where it's like you, for me, I'm also, Relatively new to this space, even though I've been doing this for about a year ish, I had to pivot from just pure copywriting and being a consultant into learning how to leverage AI. So it definitely is a journey and also finding, you know, product market fit, learning if people are looking for AI solutions or just consulting work. That has always been You know, an interesting juggle and a journey for me. And I'm sure many other founders and entrepreneurs out there. and of course, I think Tom, it will be great if you could share a bit more about SEOwind and how people can get started on your platform.

Ton Winter:

So at SEOwind, what we focus on is doing the research and actually analyzing all the SEO data that we need to create a brief and then prompt AI for you to write the article piece by piece. So we go to the SERPs, like we automate the whole process of going to the SERPs. Also understanding the keywords that you should go after. we ask you as a special matter expert, what do you think? What are the crucial parts like that you should take into account when writing this article, so we then later can parse it, divide it into specific thoughts and like align it with the outline. So actually say, so yeah, I like looking at our special letter expert things, this and this about this part when you'll be writing that. Remember to submit it, go with additional statistics and data so it will not hallucinate and make things up. We are really explicit with the things that we want. and also when AA writing the article later on, we are able to find internal links based on, Google Search Console data. So we try to simplify the whole process of researching, and gathering the data. When AI writing for every 3000 words that we write, more or less, we feed it with about 10, 000 words in the context. So there's like almost three, more than three times in what we take out, because this is the context that you need for AI writing. so we try to simplify it. Like, of course you can do it in ChatGPT. That's doable. I know how to do it, but if I'm thinking about writing 10, 000 words in the prompt and writing 3000 words of the article from scratch, I'm going with the article to be honest. Like I will not write the prompt, even though that I exactly know how it's structured because like, this is so much data that we need to compile from different sources and feed it to AI. Right. Yeah. Digestible way that it's from human perspective. It's just not doable in my opinion, at least, like I would not go with that. so this is what we try to do. We deliver you a tool that helps you out to specify exactly what you want to write about, gives you the data and then passes it to AI for you to write the article. and then in my opinion, when we write the article, it's like 70 to 80 percent ready. As an editor, go through it, add your own input into that, make it better, make it helpful, and then rank for the keywords that you want. if you want to try it out, like just go to SEOwind and you'll find a register button there, play around. If you want to learn more about it, like always, you can find me on LinkedIn, ask me. Anything you want, would love to chat with you.

Audrey Chia:

Thank you so much for that, Tom. And thank you for joining us today. It was a real pleasure having you on the show and thank you listeners for tuning in. So don't forget to subscribe to the AI Marketers Podcast and hit the bell for more exciting episodes, just like this one. Thank you. And see you soon.

Ton Winter:

Thanks a lot. Cheers. Bye bye.