Citation rich AI writer for SEO friendly content writing.

Jim James: 0:00

Hello and welcome today to Fabian Langer who's joining me from Germany. Fabian, welcome to the show.

Fabian Langer: 0:06

Thank you, Jim. Thank you for having me.

Jim James: 0:08

Now Fabian, you're on the show is the inventor of a great app called AI-writer.com. So can you just tell us how do you help to entrepreneurs to get noticed?

Fabian Langer: 0:20

Well, it's actually quite simple. If you want to get noticed, usually you need a, you need to be found on Google. How do you get found on Google? By writing content that Google is able to find on your sites so people can find you. And that's how the AI writer then can, can help you. It can help you to create this content more efficiently, more SEO friendly. So that's the, that's the whole point.

Jim James: 0:43

Okay. Now, you make it sound simple enough, but I've used AI writer and it's ai-writer.com is the website. How has it, how has it different? I've got some of my own opinions, but can you, as the founder, as the creator, explain how is it different to all of the other AI writers that have cropping out?

Fabian Langer: 1:04

So the whole approach is a bit different and what you, as the end user see it, you will get citations. So everything the AI is writing can be traced back to the original source. And that's my main difference because most of the other AI writing tools, they just make stuff up on the fly. They pull all the knowledge from one length, from one AI and the AI just keeps on writing. And with my tool, you can verify everything that is written by going to the source where the information comes from. So that's the my difference to the other tools. So you, you can trust basically, you can trust what the AI writer is writing.

Jim James: 1:44

Okay. Yeah. And I have to say I've found that really useful because especially if you are filing things for journalists, for stories.

Fabian Langer: 1:51

Yes, absolutely. If you have a perfect art people, but you cannot check what is written there. Can you even use it on a website? In my opinion, you can, because you do never know, if it's true. Only if you you're an expert in the domain, but then usually you don't need an AI writer to write for you because you can just write it yourself very quickly.

Jim James: 2:12

So Fabian, for those people that aren't familiar with AI writing, can you just take us through the steps of what one would need to do on your platform in order to get an article?

Fabian Langer: 2:24

Well, the only thing you need is your headline, but the topic you want to write about, and then the AI does the same thing that you would do as a, as a human if you start writing about a topic you're not familiar with. It goes to Google and then it puts in the search query and that gets the sources and it reads up on the topic and then it creates an article. But the AI can do it a lot more efficient than you can. It takes the first 100 results and it goes through all these information. And then it tries to get the quin-, the essence of the topic for you. So it tries to write a final article.. There was a lot of research in that, but it's, that's not, that's far from perfect. What it really does well, is it, if it gives, it gives you a, some kind of brainstorming, it helps you do research faster. So you you're up to speed about the new topic a lot faster than if you would go there and just read all the articles and all this. So that's, that's what, so you give me a headline and you'll get back an article. So it's a very simple process and this article is basically a draft. It is helping you do to get up, to speed faster. So you have something to start from, not from scratch. That's that's the main selling point.

Jim James: 3:40

Okay. And so the, the idea and I like that it saves you a lot of time of scrolling, becoming a subject matter expert, especially if you're in a PR firm or a writer, a career writer. How much, if you like alignment, can you help someone get with say the corporate view on a topic? Because often, you know, companies ask someone to write, but actually they have a perspective, for example, on, in the environment or 5G. Fabian, how, how can AI-writing help with that?

Fabian Langer: 4:11

You mean, how you can help with getting someone who's not knowledgeable up to.

Jim James: 4:16

If you've got a company that wants to write about a topic from their own perspective, so they have an attitude towards something, let's say CSR.

Fabian Langer: 4:26

Well, that's, that's a hard thing to, to get right, to be honest. So if the company has a lot of stuff published already and it, it can be fed into the AI as source material, but in the end it's, as I said, it starts all with a Google search. So if you, you can quip this search, if you want to write very favorable about the topic, for example. You need to adjust the query so it yields source material then writes in the, in the perspective you intend to. So it's a bit like Googling, it's a, it's an art form. You need to know what the query needs to look like, you know. So it's a bit of trial and error, but usually you get a feeling for pretty fast.

Jim James: 5:13

Okay. Yeah. As you say, you get to a draft don't you have like a synopsis for a topic.

Fabian Langer: 5:18

Yes.

Jim James: 5:18

And so what guidance do people have on AI writer to help them narrow down the headline that they should be writing in order to get an article that's worth having.

Fabian Langer: 5:31

At the moment, so then there is a very new feature, it's still in early beta testing. It it's called a, the Keyboard Reports. So the tool can, can go out to Google and if you, if you type in a topic in Google, it's it does some suggestions for you. So you know what other people are searching in the, in this field and it takes all these suggestions. And then it analyzes who is ranking for these keywords? How hard will it be to rank for this keyword? How much traffic is this keyword? So can analyze all the different subtopics in your domain and see which one you're able to rank in Google. So, and how much traffic these are getting. And by analyzing these, it can tell you what to write about to get the most out of your time, the most traffic. But that's a very new feature currently in the test, test fastered and can tell you if it's perfect already, or if it still needs a lot of work.

Jim James: 6:30

I look forward to trialing that. And then Fabian then in terms of the, the article linked, for example. Sometimes people want to do a short blog post 500 words. Sometimes it's a long form article for a contributed piece to a magazine of a thousand. How does AI writer help to control length, for example?

Fabian Langer: 6:51

That's, that's a hard, a hard point. A lot of people ask for it and it is, it is quite hard to get right, to be honest. So at the moment that the AI decides itself, so if it finds a lot of information that writes 1000, 1,500 words the maximum, I got from experiments. After that, it's just the quality degrades, because there is not enough information, usually from one search to write more. And if you, if the AI decides it does not find enough sources to write a lot of texts that just cuts it short. So you do not have a, a saying in that because the AI is framed in a specific way to maximize quality. And, yeah, but that so if, for example, if you, if you train it to write a thousand words and then the user says, 'okay, I just want 500'. It's not as simple as just to cut it. You would have to train it to know what is the most important 500 words. So it's harder than you think.

Jim James: 7:52

Yeah. There's very interesting actually, that it's having to synthesize that information. And how does it decide what is a value you mentioned quality of that content

Fabian Langer: 8:03

Yes, I trained it on, on hundreds of thousands of articles in the past. So it basically that's the, the, the AI part. So it, it learned it and I it's a black box for me, even as a developer, it's a give it a lot of information about the, the sentences and how they are connected and how they look like and all this kind of stuff. And then there is a huge model that learned it basically. So I cannot even tell you, even if I want it to. That's the AI part, the black box, the magic, basically.

Jim James: 8:35

Well, does it have a name then? Do you give it a name, Fabian? If it's sort of got its own personality like this.

Fabian Langer: 8:41

And not yet, but I'm open to suggestions.

Jim James: 8:44

Okay. So you put in a, if you like your sentence, and then you get back the content and you can use that cited. What about plagiarism, Fabian? Is there a control because I know plagiarism is a big issue if you use AI.

Fabian Langer: 9:01

Yes. Yes. Yes, absolutely. So, first of all, there is the fair use laws, American for example. You can use like 10% of a source, it's still fine. So that's the first, the first cutoff that I will never use more than 10% or so from a source. But on top of that, everything that the AI write, as I said, I am using sources, so there is potential for plagiarism, but everything that comes out in the end is heavily rewritten. So there is no nothing that is just copied from anywhere. It's always written in the, in the, the own words of the AI. And there is a huge focus on in the research that there is no, no plagiarism in the, in the final article. So usually if you, if you take a very good plagiarism checking to, like Copyscape. It is only able to find like 5%, 5% of the whole article. It can trace back to any source at all. So it's very, very, heavily unique in the end. Yes.

Jim James: 10:07

Okay, great. 'Cause you don't want to find yourself being shut down and having your article blocked or even I think YouTube and so on. Now we're checking that in the titles, aren't there in the, in the show notes and stuff.

Fabian Langer: 10:17

Yes, I, I'm, I'm using the, my articles on my, on my own website as well. So I, I have trust in what the, the AI writes that it does not bring you any, any problems. So, yes.

Jim James: 10:30

Then tell me, what's the sort of pricing model for ai-writer.com? Because there are different ways aren't there that people can charge for this.

Fabian Langer: 10:41

It starts very, very cheap, $29 a month. You can get, the first package and basically it goes all the way up. So there are people that, that need thousands of articles every, every month they can get a huge a package, but usually you pay me on a monthly basis on a subscription basis. And yes, that's basically the pricing model. You get, yes, you can get up to...

Jim James: 11:09

You, get you get it. Sorry. Then if I'm not mistaken, though, it's not as sort of a, a fixed price for all you can eat, is it, there's like an amount of storage or something that you have as a model. Do you want to just to explain that?

Fabian Langer: 11:20

You can get a 40 articles for the $29 subscription. So you can browse for the articles. It's very cheap compared to what you spend on time to write an article yourself.

Jim James: 11:33

Yeah. So it's less than $5 an article actually for a thousand words issues.

Fabian Langer: 11:38

Less than 50 cent an article.

Jim James: 11:39

Oh, less than 50 cents. Yeah, my math mental arithmetic. I need your AI to help me, I think, with, with mental arithmetic.

Fabian Langer: 11:46

So it's and it usually saves you around 20 to 30 to 40% of your time. So it's price point wise, it's very, very affordable.

Jim James: 11:56

Absolutely. Now, what about language Fabian? Is it across only English or, I mean you're German, or they live in Dubai, can it do, you know, Arabic for example, or is it all in English?

Fabian Langer: 12:07

Oh, only English because the, in these kind of AI models that, that work for language, you usually have to train all the models from scratch again to get into another language. And in English, you have the big advantage that you have a lot of data, you have a lot of so-, of open source libraries, a lot of open source models that you can build upon. So for example, it takes, it costs hundreds of thousands of dollars to train some models from scratch. But there is research and open source models that you can build upon in English that is not available, for example, in the internet. So it's all a matter of. And also, right, this is a one person company. So I need to focus on the, by far biggest market. It's just a matter of resources and their allocation.

Jim James: 12:59

No, that's very interesting though.

Fabian Langer: 13:01

I have a lot of customers actually that, that use, some translation, like Debell or Google translator on top of the articles to get it into a target language. And they, they are pretty good at the, so the last years they, they became pretty, pretty good in translating stuff.

Jim James: 13:19

Yeah. So that's interesting. So you'd write it English, although I guess if you're topic was in a local language domain that may prove harder. Right. But,

Fabian Langer: 13:26

Yes, unfortunately.

Jim James: 13:28

But that's okay. So, tell us about you though. You mentioned that you're a one man entrepreneur, but I've looked at the product and use the product. And it's really impressive.

Fabian Langer: 13:36

Thank you.

Jim James: 13:37

How are you getting yourself noticed as an entrepreneur?

Fabian Langer: 13:42

Well, I do the same as a, as so, so I'm using SEO myself, so that's the whole thing. Most of my traffic comes, comes from Google directly. I use my own AI written articles, I use my keyboard research and that's basically my, the main source of, of how I get noticed by people that are looking for a solution like, like mine. As my AI they go to Google first, then they find me, they try it. And I have a very easy process of getting into the trial to see if it works for you. And that's basically my, my, my marketing, if you can say so it's SEO focused.

Jim James: 14:19

Okay, well that, but that really works. That's how I found you as well.

Fabian Langer: 14:22

Yes.

Jim James: 14:24

What about for people that just want short snippets, like headlines, for example, there are some tools like headline analyzer. How are you helping people to make shorter posts or even tweak?

Fabian Langer: 14:35

At the moment, there is unfortunately, no solution I can provide in this. Because as you said, this, this small kind of stuff, there are a lot of other startups that, that for example, use GPT-3 and that works quite well for this small kind of, you don't need my technology for that. You can cite anything. You that's just, yeah, nothing that I'm actively researching.

Jim James: 14:58

Oh, I see. So GPT-3 is this AI, sort of, almost available tool that I've heard some other entrepreneurs using.

Fabian Langer: 15:06

It's an API everybody can tap into it and get his own AI writing started up basically in a few days. Do you want to, but it's, that's, that's a solution that just makes stuff up from thin air. As I said, that's nothing you can use on your website because you can never trust it. And it makes a lot of stuff that are, that's just false. You can never trust that. You need to fact check your own AI, which is a lot of work. So my opinion, it's useless.

Jim James: 15:33

You know, that's a really good point because I used one AI writer and the grammar was so shocking that it was taking me longer to, you know, correct it because it creates a huge volume very quickly.

Fabian Langer: 15:48

Yeah.

Jim James: 15:49

And that was really, really a problem actually. With, with the AI, with the content that comes out from your proprietary platform, then Fabian, is there some kind of a grammar check, a how does that work for people?

Fabian Langer: 16:01

There is no grammar check about, there is a model that is trained to write English. That's the same thing as by choosing what, what information to put into it, there is a huge black box, I tried my best to print it. I'm still working on it all the time. So it's getting better and better. But then the end, it's never perfect. I mean, you might've seen it. But as I said that, that's not the main reason to use a AI-writer to get something out that's perfect but something to help you be more effective in perfect writing.

Jim James: 16:33

Yeah, so. Yeah, so what I found was I used it to create the first draft and like do the resource investigation and then add in my own quotes and localize, for example, that content. That seemed like a really good function, the way to work that. Okay. And what's in the, in the future for you then, Fabian, for ai-writer.

Fabian Langer: 16:55

Well, it's basically, I'm, I'm always working on the models the way I want to make it better and better every, every few months. So that's my, it's the only product I'm really working on and it's the one-on-one to get perfect. So perhaps in the future you can just take it and post it on the website and it's perfect, that's my goal. Well, that's a really hard goal though. I don't know if it's achievable.

Jim James: 17:17

Well, I think that for now what you're offering and for the price point, it's a really handy and intelligent tool. I personally loved the citations and that seems like a real point of difference, which is so keen to speak to you because, it means that if I supply that to a client or to a media, it shows that I've actually got somewhere to go back to. Fabian, if you want to come back to you, how can they find you?

Fabian Langer: 17:40

They can find me on my website AI-writer.com, test the tool, and if you want to drop me an email, you find my email address there as well. As I said, one person company so everything reaches me and just reach out by name.

Jim James: 17:54

But as an entrepreneur, this leveraging, you know, AI, I'm really impressed with the product you produced, Fabian. So, Fabian Langer, thanks so much for joining me today from Germany. Part-time Germany, part-time Dubai. Thank you so much.

Fabian Langer: 18:05

Mostly Dubai. Yes. Thank you very much for having me for my, for the opportunity to get my, get heard. Get noticed, right?

Jim James: 18:14

To get noticed. Listen to Fabian Langer, who's the creator and entrepreneur behind the AI-writer platform. And I'll include all his details in the show notes.