Ryan Law, Director of Content Marketing at Ahrefs, shares 4 AI prompt tips for content marketers.
AI can transform you into a super-marketer, especially you create and produce content.
Don’t believe me?
Ryan Law, the Director of Content Marketing at Ahrefs, has seen it with his work in content marketing.
"There is a bunch of unskilled busy work that we find all of ourselves doing all day that we can actually just hand off for the time being, and I think we should do that today."
Today, Ryan will be sharing AI prompts he uses to speed up content creation and distribution. It’ll surely unlock the inner super-marketer inside of you.
In this Marketing Powerups episode, you'll learn:
- How AI can transform the way you approach content marketing.
- The role of AI in increasing productivity and reducing unskilled busy work.
- The importance of authenticity and credibility in content creation in the age of AI.
- How finding trusted advisors accelerated Ryan's career.
When you're in scale-up mode, and you have KPIs to hit, the pressure is on to deliver demos and signups.
And it's a lot to handle: demand gen, email sequences, revenue ops, and more! That’s where 42/Agency, founded by my friend Kamil Rextin, can help you.
They’re a strategic partner that’s helped B2B SaaS companies like ProfitWell, Teamwork, Sprout Social and Hubdoc build a predictable revenue engine.
If you’re looking for performance experts and creatives to solve your marketing problems at a fraction of the cost of in-house, look no further.
Go to https://www.42agency.com/ to talk to a strategist to learn how you can build a high-efficiency revenue engine now.
⭐️ Ryan Law's Tips for Using AI in Content Marketing
Since the release of ChatGPT, Ryan Law and the Animalz team have been experimenting with content marketers using AI to assist in content creation. Here are a few ways to use AI content effectively:
1. Be selective about what you use AI for. 🔎
AI models have limitations and are more effective in certain use cases than others. For example, any article requiring abundant concrete data usually results in outdated or outright fabricated statistics in AI’s draft.
Because generative AI is learning from and emulating all the existing blog posts out there, the ones that are already in its data set, those same problems you will find carried through into the output it generates because that is what a B2B blog post looks like. It'll all turn out looking the same.
One thing ChatGPT and other AI tools are good at is transforming information from one format into another. For instance, Ryan would take the automated transcript from YouTube for his videos full of grammatical mistakes, plug it into ChatGPT to make it more readable, and turn it into a blog article.
2. Front-load the article structure. 📝
If you are going to use AI to create blog articles, it's best you provide the overall structure of the article in your AI prompt.
"The structure of an article is one of the most important determinants of its success. Basically a lot of writers obsess over line edits and the construction of individual sentences. And what happens is they forget that actually the way they've written something is probably just a reflection of their own thought process."
Ryan suggests being very discerning and deliberate about the structure and being meticulous with what to include. Provide the headings and allot word count for each section of the article to emphasize the importance of one point over another.
3. Inject "information gain." 🧠
The best articles bring something new to a topic: original data, a useful opinion, a practical experience, or an explanatory framework. Google calls this "information gain" in a patent they filed in 2020.
"Generative AI struggles to provide any information gain. It functions much like a copycat content machine: creating new writing based largely on the existing literature on a given topic."
Our job as content marketers is to seed the generated content with information gain, adding real-life examples, customer stories, accurate data, and product mentions.
4. Review everything AI produces. 🤔
The biggest issue with generative AI is that it lies. Worst yet, it doesn't tell you when it's lying to you!
"For generative AI and Large Language Models (LLMs), they are not created or designed to tell the truth. That is just not part of their design. That's not why they exist. They are created to sound intelligible but not truthful."
It's why you have He further explains that AI will generate content based on the context provided, but it does not fact-check the information it generates. Therefore, it's crucial to have a human in the loop to review the content and ensure its accuracy.
Free powerups cheatsheet
🎉 About Ryan Law
Ryan Law is the Director of Content Marketing at Ahrefs. Throughout his versatile career, he has thrived in roles ranging from writer and content strategist to VP and CMO, even co-founding a marketing agency. He notably spearheaded growth at Animalz as it scaled to 130 full-time employees, generating an impressive ~2 million pageviews for their blog.
Ryan's reputation for knowledge sharing is evident through his webinars and workshops conducted in partnership with leading companies such as Andreessen Horowitz, Writer, Drift, Clearscope, Wynter, and BrightonSEO.
Outside of his professional endeavors, Ryan is a multi-talented individual. He is the author of two novels and the host of the popular Ash Tales podcast. His creative pursuits also include landscape photography and showcasing his musical talent as the guitarist for The Schrödinger Effect.
🕰️ Timestamps and transcript
- [00:00:00] How AI Transforms Content Marketing with Ryan Law
- [00:01:10] The Future of SEO and Thought Leadership in a World of AI Content Creation
- [00:09:45] The Future of AI in Content Marketing
- [00:12:26] The Importance of Thought Leadership Content in Content Marketing
- [00:15:32] Creating Thought Leadership Content Through Customer and Prospect Insights
- [00:18:12] 42 Agency - My Number One Recommended Growth Agency
- [00:18:58] Ahrefs Free Webmaster Tools
- [00:19:44] Discussion with Ryan Law on AI in content creation
- [00:21:07] Exploring the Role and Challenges of Generative AI in Content Creation
- [00:24:54] Ryan Law on Using AI in Content Writing
- [00:30:09] Discussing AI and Tone in Content Creation
- [00:33:02] Leveraging AI for content creation with Ryan Law
- [00:36:56] Transforming YouTube Transcripts with GPT-3: An Insightful Interview with Ryan Law
- [00:38:44] Career Acceleration: The Power of Content and Opinions
- [00:42:03] Overcoming Fear of Sharing Opinions on the Internet
- [00:45:30] Building a Trusted Network of Advisors in Content Marketing
- [00:47:11] Ryan Law on Doing Hard Things in Marketing, Standing Out, and Becoming Successful
[00:00:00] Ramli John: AI can transform you into a supermarketer, especially if you create and produce content. Don't believe me? Ryan Law, the former CMO of Animals, has seen it with his work in content marketing. [00:00:10] Ryan Law: There is a bunch of unskilled busy work that we find all of ourselves doing all day that we can actually just hand off for the time being, and I think we should do that today. [00:00:19] Ramli John: Ryan shares his AI prompts he uses to speed up his content creation and distribution shoe really unlock the inner supermarketer inside of you. In this marketing Pops episode, you learned first, how AI can transform the way you approach content marketing. Second, the role of AI in increasing productivity and reducing unskilled busy work. Third, the importance of authenticity and credibility in content creation in the age of AI. And number four, how finding trusted advisors accelerated Riot's career. Before we start, I've created a free Power sheet sheet to help you apply Riot's AI prompts for content marketing right away. Can download that at marketing powerups.com or find the link in the show notes and description. You ready? Let's go.
[00:01:10] The Future of SEO and Thought Leadership in a World of AI Content Creation
[00:01:10] Ramli John: Today we're going to be talking about something near and dear to your heart content with AI chat and thought leadership. I'm curious what your take is on what the future of SEO and chat GPT looks like in terms of content. I know you've been writing a ton of posts about it, and you've talked about what that could potentially look like for us. What are the implications of that? One of your viral posts on LinkedIn, which I'll link in the description, talks about in the post SEO, post chat GPT world, this is what SEO and content can look like for my listeners. What does that look like, particularly some of the implications you already foresee? [00:01:57] Ryan Law: Very good question. And all of the writing is basically just to try and work out for myself, like, what the heck is going to happen? We make our living doing this kind of thing, and suddenly this crazy disruptive technology has appeared that can do all the stuff that I can do, but in some ways so much better. I can't go out and read Wikipedia and all that. So a big part of what we've been trying to do is just think through second order impact of AI content because there's some stuff that's really obvious that will happen. It's now really easy to publish a ton of words on a page because you have a tool that is basically like a freemium, SaaS product that can do that. Very intelligibly, very coherently at the click of a button. But assuming that is the case, what happens afterwards? I think of search as an ecosystem. Not only are we trying to create things for ourselves and generate traffic for our companies, we're also doing that alongside other companies and their efforts and the things that we do impact what other people have to do. And it's all this kind of delicate equilibrium. And I think, yeah, first and foremost, we are going to see such a big influx of new search content. So many companies have built their growth on organic search because it compounds over time. It becomes more affordable over time as well. And suddenly you can actually create a very legible SEO optimized post through a tool like Jasper or Writer or copy AI or something like that. And it's as good as I used to be when I was like 20 years old and mashing stuff and cramming keywords into an article and hoping for the best. It's at least as good as that and possibly better. [00:03:36] Ramli John: Yeah. Which is like absolutely. I mean, I've noticed marketers take two positions. First is like actual awe and like, curiosity where they play around with it. On the other hand, there's this terminator, it's like it's the end of the world. They take you over the world skynet. You have this really great video, actually, that I really love about negative visualization. And you talk about how plan for the worst case, and you start thinking about it knowing that this thing is disruptive. What are the plans? You actually listed a few plans that marketers can take in terms of like, hey, here's what you should consider. What are some things that you've been thinking about for yourself? About how do you make sure that we remain relevant, especially we're so focused on what you just talked about around content and SEO. [00:04:39] Ryan Law: I'm both of those people. I vacillate between those two different states. I am awed and amazed by it. And other days I wake up and I think, oh man, I wish they'd put the genie back in the bottle because I'm a writer. I'm in trouble here. I got to do something. There is a lot that can be done even now. Even Pragmatically, even stuff that is useful given the current state of content marketing, and not necessarily only useful in the event of some crazy SEO apocalypse kind of thing. For one, there's a lot of marginal productivity gains that I think can be had through using AI. You've talked about this, actually, in terms of your podcast, the production there, there is a bunch of unskilled busy work that we find all of ourselves doing all day that we can actually just hand off for the time being. And I think we should do that because then that frees us up for the stuff that only we can actually do. And that is like higher leverage strategy and second order thinking because there are plenty of things that generative AI is just never going to be able to do because it's not within the remit of how it was built. It is designed to create legible pros based on whatever inputs you give it. It can't go out and actually do those things. It can't come to you and say like, hey, I lived this experience. I did this thing I learned from it. Here's what worked. It can tell you a story about doing that, but words on the page don't. That's not everything that matters. Also the underlying experience, the credibility and authenticity of the person that's sharing it does matter. So I think in addition to thinking about how you can publish more words on a page, how can you increase the credibility and authenticity of the things you're writing? Like what is it that you have experienced only you can talk about what do you know that other people don't because you've done firsthand? What do your friends and your network know? What problems have they solved? I think making that the kind of core foundation layer of content marketing, whether you're doing SEO or sales enablement or thought leadership, that is a really great way to continually and always future proof against what generative AI can do because you'll just never be able to do that stuff. [00:06:53] Ramli John: You actually talked a little bit about this in an article that you wrote around. You call this information gain where your AI can write approach for you. But all that stories and experiences, that's information gain that it can't add and it's kind of veering, I guess, a little bit into. Would you consider that thought leadership? I know you have his course of thought leadership, but there's this gray area, it's become a buzzword. But all that experience and information that is unique to your experience, you add that to this article and it makes it more rich, essentially. [00:07:35] Ryan Law: Yeah, absolutely. I wish I could claim credit for the idea of information game. I'm not that smart, unfortunately. But it comes from like a Google patent that came out years and years ago, one of many speculative things, Google patents that maybe it'll make its way into how the algorithm functions, maybe not. But they were basically trying to solve a problem that we all face every day, which is you look at most search results and all the content there contains the same information. It's all like a basic remix of one another. Because that's kind of how SEO functions in really competitive searches, isn't it? Ten articles performing well, how do you outperform them? You take all the best bits of those ten and you squish them together into your one article and then you publish that. Obviously, if everyone does that, then all the content looks the same. That's bad for you as the reader because doesn't really matter which article you click, you're going to get the same information. And it's bad for Google because they're serving a bunch of really homogeneous search results. So the idea of information gain was, can they algorithmically reward content that brings something new, some new facet of the topic, some new information, some new perspective hasn't been covered and actually can search reward that and actually become an incentive for better rankings. And as you say that there's plenty of ways to do that your own personal opinion. Experience is fundamentally something that you can always add in addition to the existing search results. Because nobody else is you. Nobody else has done the things you have done, build the company you've built, worked at the companies you've worked at, that kind of thing. So I think thought leadership in terms of sharing stuff that you've done and you've solved and the unique experiences you have has a totally valid strategy for better search performance, adding value above AI content, and just generally making stuff that's a bit more interesting than what people have already published. [00:09:36] Ramli John: When you said around like, there's nobody like you, it reminded me of Dr. Seuss poem, where today there's nowhere you are than you.
[00:09:45] The Future of AI in Content Marketing
[00:09:45] Ramli John: So that's such a good point there that do you see a world where it's going to be a race to the bottom? Because you're talking about AI being able to produce a ton of things that look optimized for SEO right now, and the barrier is lower and really like what will bubble up as cream of the crop? Is what you're talking about. Those uniqueness that you can add to it that makes it totally different is exactly what I'm hearing you said. [00:10:21] Ryan Law: Yeah, exactly that. I think any kind of growth strategy works at a point in time because the more you do it, the more popular it becomes, the more you attract other people that want to do the same thing. And then you get diminishing returns from that. That's true of anything. It's like a temporary form of arbitrage. And I think what we're seeing here, there's already been a race to the bottom for organic search content for the longest time. Everyone is doing it now. I started out like ten years ago, a content agency I co founded, and we had to persuade people that content marketing was a valid thing to do. And you don't have to do that anymore. Everyone knows that it already is. And it's basically become easier and easier to publish functional search content. More people know how to do it. More people can do it cheaper amount than they've ever been able to do before. And we've reached a point now where you can do it basically for free. And for the naysayers among us, if you take the average SEO article published by a person and compare it to one that you can spin up in a tool like Writer or Jasper, it's as good. I honestly see no, in terms of the ability to rank, I think it is as good as the average thing out there. [00:11:30] Ramli John: Wow. [00:11:31] Ryan Law: And obviously anyone can access these tools. It doesn't have to be big enterprise company with a big budget to hire a big content team. You can be mom and pop shop solopreneur and you can suddenly generate like 5100 thousand articles a month if you want to do so. I think, yeah, it's going to be pretty crazy the next few months. And when you can generate content that quickly and that cheaply, you may as well just target every keyword. Why bother being selective about it? There's actually no need to in that case. So if you were hesitant before, maybe you'll just publish like 50, 60 vaguely related keywords every single that's absolutely crazy. [00:12:10] Ramli John: And I guess that's the part where, man, this is a scary world. There's going to be a ton of really content that's optimized for SEO that's written by AI.
[00:12:26] The Importance of Thought Leadership Content in Content Marketing
[00:12:26] Ramli John: Also what I'm hearing is would you say that tall leadership is the weight of the future? That sounds so wrong, but would you say that's how you'll win with how really to make your content stand out is what I'm hearing? Or did I hear hear that incorrectly? [00:12:50] Ryan Law: So I think I get a bit of criticism from people when I talk about thought leadership content because they say, Ryan, that's a really stupid thing to talk about because thought leadership is a status you want to acquire. You want people to care about you and care about the ideas you have and not everyone can do that. I think that's true, but there are certain types of content that are more conducive to creating that status. Like publishing 20 how to articles about basic marketing processes. Probably not going to become a thought leader. Whereas sharing like personal experience stories, anecdotes refuting industry truisms, doing original research and data collection that is much more likely to lead to that status. And I think a lot of those activities for the time being, they are great ways to differentiate from this race to the bottom. Basically, what proprietary information can you and your company create or gain access to that nobody else can gain access to and put that in your content? You need a moat in content. Basically we're all so inured and overrun by blog posts and PDFs and downloads and webinars. You have to really incentivize people to care about what you're creating and the best way to do that is to have something that they cannot get anywhere else. It's a hard thing to do, but I think that is probably the best thing to do these days. [00:14:09] Ramli John: You talking about that just like made me realize that who is writing the piece or what company is writing the piece adds a lot. It's probably been true for a while, but it's going to be more true in the future, where there's like a writer content fit or company content fit, exactly. Where your credibility as a person, as a writer, as a creator, as a company really means a lot more for that thought leadership. Because you're embedding your experience. Exactly, is what you're saying. [00:14:42] Ryan Law: I totally agree. I love that idea so much. And we're already seeing Google, they've been talking about this for a long time. They added an extra E into their Eat acronym, didn't they? For like experience. They want to reward people that have experienced the thing they're talking about. [00:14:56] Ramli John: Interesting. [00:14:57] Ryan Law: And if you think about it, I could write an article that might be accurate about how to drive a Formula One car. I could go out, I could probably find enough information on the Internet to write that article and it'd be strictly true and strictly accurate. I'm a terrible driver. Are you going to trust my advice on that topic? No. If the same topic was published by a Formula One driver, suddenly that advice becomes much more credible. So I think the provenance of what we share as well as the actual advice itself is going to become more and more important.
[00:15:32] Creating Thought Leadership Content Through Customer and Prospect Insights
[00:15:32] Ramli John: You got me distracted with Formula One. Do you watch it? [00:15:38] Ryan Law: I picked a really bad example. I know nothing about Formula One other than that it exists. [00:15:43] Ramli John: I was just curious. I watch it on Netflix. There's like Lewis Hamilton and Max Verstappen and I always ask people who they go with. So just a random question there in terms of actually embedding those personal experiences. And you talked about going against challenging tourisms and then really embedding those modes. How would people go about doing that? Do they start with a structure and then adding on those experiences? Or would you go the other way around and actually interview? Maybe both. Let's interview people around, experts into this topic, including yourself, and then build a structure around that. [00:16:30] Ryan Law: It's a good question and it's really hard to build a process for coming up with that kind of creative ideation. That's a really hard thing to do because obviously with search content you have keywords and that's a very concrete starting point. Whereas with thought leadership and related ideas, where do you start? I think what we always did at Animals that was a great port of call was start by talking to customers and sales prospects. What are the problems they're encountering every day? Do we have an experience based on that that we can talk about? That's a really great starting point. You can get hundreds of articles out of that. Very simple app, probably. I think then the more you do that, the more you get a sense of what are the best practices. Lots of people subscribe to one of the articles, Animalz that seem to really strike a note with people was about HubSpot and how a lot of people, they look to HubSpot as this amazing example of content marketing, rightfully so, but they then try and copy what they do for their company. And you know what? HubSpot is a post IPO public company with a huge budget and actually the stuff they're doing today is not what got them to the point they're currently at. So trying to copy that as a startup precede or Series A or something, that doesn't work. But I only was able to write that because about 20 people said to me on sales calls, I love what HubSpot's doing. Let's just do what HubSpot's doing but for our industry. So some kind of like way of imbibing these experiences, these truisms, these best practices at scale from your customers and your and prospects you're talking to, you will get more content than you'd ever know what to do with, I think from doing that.
[00:18:12] Ramli John Discusses 42 Agency and Asiat's Free Webmaster Tools
[00:18:12] Ramli John: Before I continue, I want to thank the sponsor for this episode, 42 Agency. Now when you're in scale up growth mode and you have to hit your KPIs, the pressure is on to deliver demos and sign ups and it's a lot to handle. There's demand gen, email sequences, rev ops and more. And that's where 42 Agency, founded by my good friend Camille Rexton, can help you. They're a strategic partner that's helped b two b SaaS companies like Profitable, Teamwork, Sprout Social and Help Doc to build a predictable revenue engine. If you're looking for performance experts and creatives to solve your marketing growth problems today and help you build the foundations for the future, look no further. Visit 40 Twoagency.com to talk to a strategist right now to learn how you can build a high efficiency revenue engine. Thank you also to the sponsor for this episode. HS free webmaster tools. If you want to rank your website higher in search engines, you have to make sure that your website doesn't have any technical SEO issues. Because if you do, that's like trying to run a race with your shoes tied together. That's how you lose and we don't want that. Luckily, Asiat's Free Webmaster Tools can crawl up to 5000 pages to find 140 common technical SEO issues that could be holding your site back from generating valuable traffic. It can also help you find your strongest backlinks as well as analyze keywords you're ranking for and see keywords search volume and ranking difficulty. For each of those keywords, you can sign up for Free@htraps.com webmastertools or find that link in the description and show notes.
[00:19:44] Discussion with Ryan Law on AI in content creation
[00:19:44] Ramli John: Well, let's get back to the episode. In terms of another tip that I found in that article that you mentioned, one of the things that's super important is that I think this is a video you created, actually. I'm going to link it in the show notes where you said that AI lies and when it lies, it doesn't tell you that it's lying, which is like, that's true. It could be giving you completely wrong information and I feel like that talking about like, oh man, AI is going to take over our jobs. There is always going to be a need to make sure that there is not just that experience, but there is that accurateness to the pieces that we're writing and we're not putting on content that is not. I'm using the word lie, but even slightly off that somebody can tell right away this guy doesn't know what he's talking or she's talking about because they've said that completely wrong or in a different way that it's not possible. I guess I'm curious, how do you check if it's truthful or useful content, especially if, like, it's? I don't know. Like sometimes people lie. Well, and then I guess you need somebody who is an expert in that space to like, hey, that's that's off is exactly what you're leading with. That is that correct? [00:21:06] Ryan Law: Yeah.
[00:21:07] Exploring the Role and Challenges of Generative AI in Content Creation
[00:21:07] Ryan Law: So one of the things we did at Animals is we wanted to dedicate a couple of people's time to exploring how can we use this technology, how can we help companies grow with it, how can we find a place for it within the kind of traditional services we offer? How can we learn about it? And this was basically problem number one to solve if you're trying to publish a bunch of content with by very nature of how generative AI and Life language models function, they are not created or designed to tell the truth. That is just not part of their design. That's not why they exist. They are created to sound intelligible. So you feed it a bunch of related concepts and words, a prompt, whatever you want. And the challenge it is solving for itself is basically predicting what is likely to come next in terms of the context you provided based on the rules of language as it understands them from having read like billions and billions of pages of data. So at no point is there any kind of loop, feedback loop for it to go, oh, actually, I should go and check. I've written this quote from this famous CEO that sounds like it fits within the context of this blog post, but I should go and double check. That does not do that. It just makes stuff up. And if you're writing a blog post about B, two B SaaS, we tend to quote CEOs and CMOS and stuff in those blog posts. And it knows that and it will do that and it will make it up. Maybe the person it attributes it to will be real. But the actual quote it creates, for the most part, it will not be real. And you have to really hammer this into your own psyche in terms of using these tools. It is not trying to generate truthful information. It is like throwing word spaghetti at the wall and hoping interesting fragments come out of it. And the way we solve that very simply was exactly, as you say, whatever topic we're writing about, find somebody that knows it really well and that can actually understand the shiboleths of that topic. They can see when there are these subtle giveaways that something's not quite right. A great example. One of the writers we work with, lovely person, Angela at Animals. They know a lot about ecommerce. They know so much more about ecommerce than I do. And I would look at what seemed to be a fairly intelligible blog post on an ecommerce topic and I would see nothing wrong with it. Angela could pick out 20 things that were not quite right. We talked about SMS, we didn't talk about MMS and that kind of thing and you can't really replace that. So if you're going to publish this stuff without a human in the loop, you kind of have to be willing to accept people that know this stuff are not going to be convinced by it. And if that undermines the legitimacy of it, then maybe you shouldn't be publishing it in the first place. [00:23:52] Ramli John: And that ruins that whole credibility now of the brand, the company, the person writing it, who the byline is under, especially if there's like, this is off in the future would perform worse in terms of I'm not entirely sure what Search will look like in the future. We're seeing bits and pieces of it, but if it doesn't, it's not credible. [00:24:18] Ryan Law: It won't be pushed as much. It's very easy to think about content marketing in terms of aggregate results. You know, we care about traffic and month on month over month growth and it's very easy to forget that these are people at the end of that. Every page you is a person and something that sounds coherent. And you publish it and you feel good about it and you get some page views to it that may feel like you're doing the right job, but actually if everyone reading it is going, you guys don't know what they're talking about, and bouncing off the page, then is that actually going to have an impact on their business? Probably not. [00:24:53] Ramli John: That's so funny.
[00:24:54] Ryan Law on Using AI in Content Writing
[00:24:54] Ramli John: You were talking about how Animals was exploring using AI. Actually there's this article or blog post that you wrote about content cyborg and I love that because it's a little bit of alliteration and the word cyborg is just a fun word to say. And one of the tips you talked about there is like front loading your article structure. I'm curious what that means. In some sense I understand what it is like. Have your points, have your structure, and then let AI kind of fill in the blanks in between. Is that what that's about or is. [00:25:30] Ryan Law: It something else completely, exactly about that? I think from hard earned experience, the structure of an article is one of the most important determinants of its success. Basically a lot of writers obsess over line edits and the construction of individual sentences. And what happens is they forget that actually the way they've written something is probably just a reflection of their own thought process. Oh, I know nothing about this, I have to write a blog post. I'm going to Google, what is this thing? And I'm going to Google some common tips and then I'm actually going to address the actual topic, like how to do a particular thing. And what you end up with is something that probably well written, but you look at it and it's like somebody has gone from I don't understand this to actually trying to solve the problem on paper. Like you can actually see their thought process and for the most part that's not very convincing. If you're trying to attract anyone that isn't like a total newbie like you are to that subject, you have to get that off the page and be a bit more discerning about the things that they will find interesting and not the things that you found interesting as you are new to the subject. And because generative AI is basically learning from and emulating all the existing blog posts out there, the ones that are already in its data set, those same problems you will find carried through into the output it generates because that is what a B to B blog post looks like. So actually what we found was do not trust it with something so important as the structure. We have to be very discerning, very deliberate about that and we have to be meticulous with what we include interesting. So what we did is we would basically write out fairly detailed header structures for the information and even have like relative weightings for the information. Because if you got 1200 words you have to solve the problem you've set for yourself and you can't afford to meander off into different directions, which generative AI will totally do. If you left it in terms of. [00:27:31] Ramli John: Waiting, what does that mean? Do you just prompt it like, hey, spend more time in this why part or this section rather than the rest? Or how would you weigh it over another part? [00:27:44] Ryan Law: Yeah, and this is something I encourage normal human writers to do and something that I try to do. You have a topic, you work out the five things you need to cover to cover that topic. Chances are a couple of those are going to be much more important than the others. Maybe the other things are just context. You need to be very deliberate about that and needs to be reflected in how you allocate word count. You want most of the word count to go toward the thing that matters most. And obviously because it's so easy to write words with Generative AI, you could end up with like a 3000 word article very easily and that's not actually good for the reader. So exactly that. If something's less important, say in 200 words, give me a summary or an overview of this and then focus on the more important part of the topic. [00:28:29] Ramli John: Interesting. Are you giving it word? Maybe in certain situations. Are you giving it a specific word comp for a specific section because you want to emphasize that more? [00:28:40] Ryan Law: I've tried doing that. It doesn't seem to listen to me very well in most cases. There's lots of manual editing and whittling down, but saying stuff like summarize this or make this more concise, that kind of thing generally does seem to help. [00:28:56] Ramli John: That makes sense, what I'm hearing. I just want to make sure. I'm getting this right. You got the headers, the main points. Do you have a little bit of bullet points in between for each of those points? So you've already done the meaty research and you might even have quotes in there and some points and then you plug that into Chat GPD or whatever. AI, Jasper or Writer is exactly what you plug in, is that right? [00:29:26] Ryan Law: Yeah. I personally think there's a lot a human needs to bring to the table for this workflow. To actually work. Yeah. People need to have an idea. Yeah. Idea of structure, idea of key points, anything related to your product messaging that you want to communicate as well. Because not all tools have that context. Something like Writer might do because they're building with that in mind. But a lot of other tools don't, right? Yeah. And a lot of people beg the question like, this is a lot of effort, why don't you just get a person to write it? And in some cases you actually totally should get the person to write it. If you feel like you're fighting against it, it's not being worthwhile, then reevaluate that. But I think there are opportunities to improve your own workflow with this stuff.
[00:30:09] Discussing AI and Tone in Content Creation
[00:30:09] Ryan Law: That's cool. [00:30:10] Ramli John: I'm curious, what's your take? How do you craft the tone of it? I've seen where make this blog post sound like a HubSpot blog post, another one which is more fun, like make it sound like Ted Lasso, which I tried once and it was like, that sounds cool, like it's very Southern enthusiastic. But do you give it prompts in terms of how you want it to sound like or do you feed it some information about make it sound like me and you link a bunch of blog posts from the past? [00:30:44] Ryan Law: I think if you do, if you are particularly hot on a particular tone of voice or style or whatever, I've seen great results from kind of contextual prompting. So you say generate a tone of voice that's X, Y and Z. Here is an example of what that looks like. And the more you can give it examples and context, the more it can look at it and go, oh, okay, I see what you're getting at here, let me emulate that. It's kind of an interesting thing to do because I see people do things where they put their blog post in, they go like, tell me what my tone of voice is, and it pumps out a bunch of descriptors. I don't think that's hugely useful though, because telling someone that they're enthusiastic, what does that actually mean? How can you learn from it? Can do a good job at emulating particularly unique tones of voice, I think, like publications like BuzzFeed or whatever, it can get that bang on. [00:31:38] Ramli John: I think that BuzzFeed is interesting. I've never done that. It's probably like shorter words, like even a kid would understand. Are you using that kind of like have you used that in the past? Or emulate BuzzFeed or emulate HubSpot or I don't know, even emulate animals like their articles. I'm sure that's enough establishment there. [00:31:59] Ryan Law: No, I've not done that. I should do the very narcissistic thing of get it to try and copy the Animals blog and see what happens. I'm going to play around with that and see if I can learn from myself. [00:32:09] Ramli John: That's so funny. This is so crazy. This is something I've been thinking a lot about. I use Chatchy Beauty to write some of my YouTube scripts. Like I've been doing short form, like, like five minute short videos. I even asked it to like give me some puns and jokes. It would imbue like whole jokes about like, oh, this is like, I don't know, just some really corny jokes that I thought I think it's funny and it's really interesting how much you can add in terms of the output based on your prompt is exactly what's happening here. [00:32:46] Ryan Law: I think anytime you ask it to go, I've done this thing. Can you give me 20 more examples? It's so good at that. It's actually quite hard for a human brain to come up with 20 different versions of the same thing. But yeah, AI does a great job.
[00:33:02] Leveraging AI for content creation with Ryan Law
[00:33:02] Ramli John: I know we've been talking a lot about AI and it seems like you played around it more than other people. Is there any other tips that you have, any tips before we switch gears? In terms of people who are tuning in? How can they create better content as a content cyborg AI as their personal assistant that can help them actually create content that sounds and good. [00:33:30] Ryan Law: So most of the use cases I've found that have actually hugely saved me time and made me a very happy man are not necessarily writing things because I'm an okay writer, I know how to do that for the most part, but there's loads of writing adjacent stuff that I'm getting so much help with. A big part of that I think is transforming information from one format into another. So like YouTube video transcript, something I've been doing which is super fun, I record a five minute video upload it, get the automated YouTube transcript and it's terrible, it's wrong, it's got timestamps everywhere. I paste that into Chat GPT and I say, please reformat it. And it does, it pitch perfect, corrects the mistakes, capitalizes everything. And then you can say pull out the three core ideas from this and turn it into a basic article and it will do that and it's not going to win awards. But in terms of making a more accessible transcript, you can do that in seconds. That would have been really laborious and kind of sucky for me to do in the past. And I'm also Generative AI for images. I use it every single day because I am not a good designer. So actually even basic things with that mid journey, for example, are just blowing my mind. I'm creating custom illustrations for everything I publish on my personal site at the moment. Trying to get like a cohesive brand identity. And it's so easy to do. It's almost sickening how easy and how fun it is to get custom images. [00:35:06] Ramli John: In your blog. Oh, it's your blog post. Are you also doing it for I know you have this content concepts where it looks cool. You talk about weasel words and pareto content. Are those created by mid journey or really? [00:35:21] Ryan Law: Yeah, 100%. All those images. [00:35:26] Ramli John: What is your problem for that? Do you plug in some previous image and make it black and white and give me checkoff's gun, like a gun on a wall? [00:35:37] Ryan Law: Basically I think of like what is a good visual representation of this concept. That's the hard part. And years of doing that for the Animals blog have helped me get vaguely okay at doing that and think of whatever visual metaphor you want and then try and create some kind of consistent style through the prompting. So I say stuff like simple line illustration, black and white and then you the aspect ratio you want with the model you want to call on. That's really cool. There's some crazy stuff you could do around image seeding where you can actually use the same seed for every generation and theoretically get it more consistent. But that's slightly beyond my level. [00:36:13] Ramli John: This is just blowing my mind. Looks so good, the images. I'm going to link your content concepts and essays. Ryan Lawaud, me. But it is just like wow, that's all created you said mid journey, right? Because there's a few other mid journey. [00:36:29] Ryan Law: I started using it to create images for the fiction I'm writing because it is sensation, like fantasy style imagery. [00:36:36] Ramli John: Do you have a new book coming? New fantasy book coming out? [00:36:39] Ryan Law: Or like I know you some coin. I've written two so far and I'm trying to finish the trilogy. [00:36:44] Ramli John: But that's project that's super cool. I'm going to link those. I actually have bought it. I haven't started but I will read the first one for sure. [00:36:53] Ryan Law: You're making me very happy today.
[00:36:56] Transforming YouTube Transcripts with GPT-3: An Insightful Interview with Ryan Law
[00:36:56] Ramli John: I actually want to switch gears. Thank you for talking about AI and making that more useful for content. I find that really interesting, like repurposing, actually. I have a follow up question around your YouTube. One chat GPD has a limit in terms of the input. Are you using a plugin to take that? I'm not sure what they're limited in terms of word count in terms of the input, but is it fitting in when you paste the YouTube like your transcription or your caption right into thing and then you say, hey, fix this up. [00:37:31] Ryan Law: So I am hitting the upper limit with some of my longer videos. Anything less than ten minutes, the transcript actually seems to fit in a single prompt because they're constantly increasing them right and I do find if I have to paste in two parts, the result so far has been way worse than if I can provide the entire thing in one go. But they're constantly increasing the limits of the I didn't know that. [00:37:54] Ramli John: And then you just ask it, fix it. Because I have this problem where YouTube takes the caption. I'm like, my name is not Rambly. Or you voice your prompt for it. Like, fix this up and make it readable. Is that what you tell? [00:38:14] Ryan Law: I say stuff like, turn this transcript into fluid prose, remove timestamps, correct typographic errors, and in that case replace the word Rambly, like rambly and the actual words. You can be quite specific about that. And so far, yeah, it's actually been. [00:38:34] Ramli John: Pretty consistent with pros. I like that. Okay, I'm going to do that right after. Thank you for sharing that. I just got something awesome to go away from here.
[00:38:44] Career Acceleration: The Power of Content and Opinions
[00:38:44] Ramli John: I'm going to shift gears and talk about career power ups. Those are things that help you accelerate your career. I know you've been in content marketing now over a decade. You started your own agency, you worked at Animalz, and then now you really like so much experience in terms of marketing and content. I'm curious what's helped accelerate your career? It could be many things. It could be like networking or talking to people, but it could also be a hard skill, like learning how to use AI. I'm not entirely sure. [00:39:21] Ryan Law: Probably reflecting back. One of the things I'm really glad I was able to do is write stuff that voices an opinion, basically, because I think back on the early, like five, six years of my career and I was writing search content, lots of utilitarian, like how to stuff, what is posts, all that kind of thing. And it was great. I got our first company blog up to like a million page views in a year kind of thing. [00:39:47] Ramli John: Wow. [00:39:48] Ryan Law: Humble four person agency blog just on the back of search. But the thing is, nobody really cared about it. Didn't result in any new business, didn't result in any career opportunities for me or anything like that because it was just like totally utilitarian, functional content that basically left your head as soon as you solved the problem or another page. And as soon as I got to Animals and I started working with Jimmy Daley, who was heading the Animals blog before I took it over, they published Pinionaid content. Like, they actually shared opinions. They saw what was happening and they said, I agree with this or I don't agree with this. And it was really scary to begin with because you don't want to risk being wrong and there's always that risk you run. But by polarizing people, people actually remember it. If you resonate with them, they remember that idea. Maybe you validate an opinion they've held before. Even if they don't agree, they remember you for doing and as long as it's a defensible human, not monstrous opinion you've had, I think it's a great way to stand out because most content does not share an opinion. So, yeah, very lucky that I had an avenue for doing that and not really realized that was what I was doing. People come up to me now and they say, like, oh, I remember you shared this, and it totally changed my mind about something. They remember that it actually sticks in their head. [00:41:09] Ramli John: I find that interesting. I was talking to Tommy Walker yesterday from the content studio. You were on that show. I should link your I'm going to be on that show as well. But one of the things that I landed on and it told him was like, great content transforms people. It's like this content is a job. There's the whole concept of jobs to be done, and the concept is products transform people to become better version of the self. And it's true for content, too, where, like, you sharing an opinion kind of helps shape their people's. It transformed them. Exactly what you just mentioned with that person. Well, thank you for changing my moment about this, because you help change who they are, and maybe I'm reaching here, but who they'll be and what they'll decide because of what you shared there. Exactly.
[00:42:03] Overcoming Fear of Sharing Opinions on the Internet
[00:42:03] Ramli John: I'm curious, how do you get over that fear? I know for people who are like you mentioned about being wrong and sharing an opinion, and the Internet can be a very harsh, harsh place. Especially, I'm not sure, but I find sometimes Twitter can be really harsh with people who have wrong opinions. How did you get overcome that? Especially for people who are tuning in, who are like, I i don't want to share my opinion. I'm scared of being shut down right away. [00:42:39] Ryan Law: I'd not entirely overcome that because I still take criticism incredibly personally. I wish I was thicker skinned. I still find it very hard. I actually published something a couple of months back, and I ended up seeing this whole LinkedIn thread that was just dragging the idea. And I took it so personally, I could walk away from my computer for a bit and kind of introspect and reflect on what had happened. And it's a hard thing to get over, I think, having people around you whose judgment you trust, who are willing to support you in what you're doing, that is a really useful starting point. Because if you can find five people that go, you know what? This is a good idea, then even if you get some criticism out in the world, at least you have the defensibility of other people thinking it was a good idea. Like you unified around sharing that together. I was always very lucky to have that Animalz. People beat up my article ideas with me. We'd go back and forth over it, play with different framings, and they would tell me when I was wrong, because I was wrong many, many times in terms of things I've thought or said. And I guess the other part of it is just maybe start small. You don't have to be the iconic class that goes out and starts ranting about why X Company is wrong or why Y strategy is the stupidest thing you've ever heard. Yeah, something I talk about in the first course I put together is this idea of like, yes and ideation. So a really great way to get started is to find something you agree with and then add additional context to it. If somebody shares a really great opinion, you can publish your own article that goes, we agree with this concept. Here's how the practical experience we have that validates this. Or I zoomed in on one specific subtopic and went into great detail about this and that kind of de risks that existential problem of whether people will hate it or whether they'll think you're wrong and stupid terrible. [00:44:30] Ramli John: What I like about that is when you surround yourself with people you trust who are honest with you, you can almost like, workshop your opinions with them and kind of like, hey, am I totally off the balls, off the nuts here? Does this make sense or not? And them giving you feedback can help you kind of gain that confidence. To take that leap is exactly what I heard you apply to. You sharing your opinion totally. [00:45:04] Ryan Law: Although I would say there is one caveat there, which is that if you bring too many people in, it's very easy to end up kind of this committee led ideation method and end up watering everything down to appeal to the broadest cross section of people. So have some big spicy opinion and a couple of people whose judgment you trust intimately and then leave it there. Just publish it and let it fly and see what happens. [00:45:29] Ramli John: That's super.
[00:45:30] Building a Trusted Network of Advisors in Content Marketing
[00:45:30] Ramli John: How many of those people do you have? Like a handful, like five or less. I'm just curious. I think this is really great concept, not just for content, but or marketers, but like, life in general. Have those handful of trusted advisors that can give you opinion on your work or your career direction, so to speak. And that's done well for you. You said there's a point where there's too many. For me myself, I think I have three people that I trust and talk to about the growth in content and marketing in general. But is that about the same number that you have in terms of people that you chat with? [00:46:15] Ryan Law: Yeah, probably maybe a couple more. Because basically I was very lucky that when I joined Animals, the people that were running it at the time were just so much smarter than me and they all had different skills that I didn't have and they could all bring new perspective to it. So there was Andrew, who was this amazing, critical thinker and an actual legit neuroscientist who brought this entire background to it. There was Jimmy who was just like the most legendary marketer ever. Devin, who became CEO at Animals and she's like this huge visionary that always challenged me to think bigger about what we were doing. And like Walter who founded the company and he smartest dude I've ever met. He would tell me like, right, this is stupid. Let's work through and make this better. And it would always be better as a result of that, right? So between that little like, brain trust and those different perspectives, I learned a huge amount. [00:47:09] Ramli John: That's super cool. Thank you for sharing that.
[00:47:11] Ryan Law on Doing Hard Things in Marketing, Standing Out, and Becoming Successful
[00:47:11] Ramli John: One final question. If you can send a message to younger version of you, of Brian, what advice would you give yourself who might be like just freelancing starting out, kind of figure out what he wants to do, what would be that message you can send across the time? And once again, it could be something we've already talked about or something that you haven't shared yet. [00:47:35] Ryan Law: To your younger self, I'd probably say two things. First up, the willingness to do hard things is a really powerful moat. If you are willing to do something that other people won't do, that is a great way to set yourself up. The example I had is of like when I was starting out in content marketing, I was the dude that would endlessly grind out 10,000 word articles and people didn't want to do that because it sucked. I hated it, I didn't enjoy it. But it became this thing, this shtick that I could talk about and kind of vaguely get known for and I'd get good results from it all because I was willing to sit down and grind that out over the course of a couple of days. And if something's quite easy to do, everyone will do it and it's very hard to stand out and differentiate from those. So maybe you can be the person that goes and interviews 20 people for the next article you write. That kind of willingness to do hard things is such a powerful moat. And I guess something kind of related to that is don't compete on the same terms as other people. If you can avoid that, think about like job applications or whatever. If you go through the standard process of application and CV and you are relinquishing a lot of the control over that process, you have to be very, very lucky for that to work out in your favor. Especially in like a crazy job market now. But if you can be the person that's building their own content and actually building their own audience and showing people that you have the skills and not just talking about it, that's amazing. Or maybe you're the one that's networking and trying to actually get in touch with the people that might eventually hire you and running all these calls or whatever. If you can again, do stuff that other people aren't willing to do and that can really set yourself up for success. [00:49:23] Ramli John: That's so good. That kind of like ties back to everything we just talked about. AI makes everything easier, but if you do the hard stuff. [00:49:32] Ryan Law: We came full circle. He did it. That's beautiful. I love it. [00:49:36] Ramli John: It was all planned out. We have this all scripted. We love this conversation with Ryan. He's such an awesome dude. And learn more about Ryan, you can follow him on LinkedIn and Twitter. He also has courses on toddlership at Ryan Lawadpodia.com. All those links are in the show. Notes in Description thanks to Ryan for being on the show. If you enjoyed this episode, you'd love the Marketing Powerups newsletter. Share the actual takeaways and break down the frameworks of world class marketers go Marketing Powerups.com subscribe and you'll instantly unlock the three best frameworks that top marketers use, hit their KPS consistently, and wow their colleagues. I want to say thank you to you for listening and please like and follow Marketing Powerups on YouTube, Apple Podcast and Spotify. If you feel extra generous, kind of leave a review on Apple podcast and Spotify and leave a comment on YouTube. Goes a long way in others finding out about Marketing Powerups. Thanks to Mary Soldon for creating the artwork and design, and thank you to Fisal Tygo for editing the intro video. Of course, thank you for listening. So for now, have a powered update. Marketing Powerups until the next episode.