So What? Marketing Analytics and Insights Live
airs every Thursday at 1 pm EST.
You can watch on YouTube Live. Be sure to subscribe and follow so you never miss an episode!
In this week’s episode of So What? we focus on AI assisted content marketing. We talk through a head-to-head tool review of AI content tools, use cases, and how to build a process around AI for your content marketing. Featuring: open.ai, lately.ai, gocharlie.ai, openrank.io, and copy.ai. Catch the replay here:
In this episode you’ll learn:
- a head to head tool review of AI content tools
- use cases for using AI-assisted content writing tools
- how to build a process around AI for your content marketing
Upcoming Episodes:
- TBD
Have a question or topic you’d like to see us cover? Reach out here: https://www.trustinsights.ai/resources/so-what-the-marketing-analytics-and-insights-show/
AI-Generated Transcript:
Katie Robbert 0:29
Well, hey everyone, Happy Thursday. Welcome to so what the marketing analytics and insights live show. I’m Katie joined by Chris and John. This week, we’re actually doing essentially sort of what Chris likes to call a bake off. But basically, what we’re doing is we’re going to be looking at different AI assisted content writing tools. Because a few weeks ago, I started this project of seeing if an AI Assisted writing tool could help me produce more content and in that really to sort of get me started with writing the content versus writing it for me. And so we’re going to be looking at five different tools and sort of going through the use cases for an AI assisted content writing tool, and how to build a process around it. So before we get into the different tools, Chris, you want to walk through a little bit of the technology behind an AI Assisted writing tool, and sort of you know where these came about?
Christopher Penn 1:29
Sure. So most of the modern writing tools these days use a technology called transformers, nothing to do with the 80s toys, which were awesome, or the Michael Bay movies, what they do is essentially take in really large amounts of writing large amounts of written text. And if you can imagine, like a spotlight traveling across a stage, that’s essentially what this this technology does is it sort of reads through text and develops probabilities like vice, say, social media, what’s the next word that comes to mind? For some people, it’s marketing, for some people, it’s analytics, and so on and so forth. You are played a game at a party where you start quoting pieces of song lyrics, right? Almost everybody knows the next word, a set of words in certain songs. That’s essentially what transformers do. They have these really, really, really large, they’re called embedding spaces. But it’s just a way of saying, here’s the probability of the next logical set of words in this moving window is going to be X or Y, or Z. And in doing so, when you ask it to create something based on a prompt, it will look for all the embeddings that are related to that thing, and then start generating text from those things. So again, thinking of like, famous song lyric lyrics that, you know, if I say it’s nine o’clock on a Saturday, and you’re 70s fan, you probably know the regular crowd shuffles in. That’s Billy Joel’s piano, man, if I started, if I said write a song about, you know, being a piano player in a bar, it would probably borrow from some of that text and generate from that. So that’s kind of how the technology works. There are limitations. The key limitations are, it’s trained on a lot of existing text. And let’s be honest, a lot of the existing text is not great, right? It’s not great writing. It’s not Pulitzer Prize winning, it’s a lot of it’s kind of, okay. And so the tools will inherently generate writing, there’s also Okay, right, it’s not going to generate a Pulitzer Prize winning content. And the second problem they have from a technological perspective is something called a tension tool called attention memory deficits. These tools don’t remember what they write, because they don’t actually understand anything. So you could type into some of them six plus seven words, six plus seven equals, and it will come up with something that has no relation to what the next word should be, which is 13. Right? Because we know, logically, that would be the next word in that sentence, but the schools don’t understand that.
Katie Robbert 4:08
So what I’m hearing you say is, there’s a good reason why my phone no longer auto corrects, to ducking because I’ve trained it on the word that I really want it to be using.
Christopher Penn 4:24
Yes, but your phone uses a different architecture uses something called an LS TM, long short term memory, as opposed to a Transformers. But yes, it’s similar in concept, sort of,
Katie Robbert 4:35
well, it’s the predictive text basically, is what it comes down to is if this if this is what I’m, if I’m reading hundreds of millions of pieces of content, and someone asks me, What is the next word and, you know, social media, then based on all of those things, that’s the predictive text it’s going to be marketing or manager or you know, whatever the thing is, And I, what I’m understanding is like, your phone essentially sort of does the same thing. It’s the natural language processing, to say, okay, based on the type of content that this individual user creates, versus the big widespread these content, writing tools, this is the next likely word that I’m going to suggest.
Christopher Penn 5:25
That’s right. And with these larger language models, you can do what is called fine tuning, where you would take for example, open AI offers this you would give you take it starting model, like the DaVinci, GPT, 3d model, and then you give it your entire corporate blog, and you can say, Okay, now, I want you to learn from this writing style and tune it so that what you create sounds more like our corporate blog, that would be an example of fine tuning it to make it a little bit more custom.
Katie Robbert 5:50
Okay, so I don’t want to get too too bogged down in the weeds of the technical details. Because what I think a lot of us myself, you know, maybe John John, you tend to be more technical than I am, is, what we’re trying to understand is, are these tools that can help us do our jobs more efficiently. You know, and so, like I said, I’m not trying to replace the task of actually writing the content, what I’m trying to do is sort of get like a head start on writing the content, because that’s what takes me the most time is actually getting the content started. And what I found, at least so far, is these tools can at least give me like a rough outline is something to react to, when I have that thing to react to, like, oh, okay, now I know exactly what to do. And it takes me a fraction of the time to finish the post start to finish. John, do you use any predictive text or content writing tools for any of the stuff that you do?
John Wall 6:54
No, I really haven’t been able to take advantage of any of that stuff. You know, I’ve started to look for blog posting and getting some content out there. But ya know, like, everything that I have is like a half list of like cutting edge topics, you know, there’s like stuff that’s not out there. And I haven’t found anything, that’s an easy enough tool for me to just throw stuff in and see what I can get out of it. And so that’s what I’m hoping we’ll dig into a little bit today, see if there’s something that should be on my radar that I could put to use without having to go through a whole ton of configuration?
Katie Robbert 7:23
Well, I think that is exactly what we’re going to be doing today. So Chris, where do we want to start?
Christopher Penn 7:29
Let’s start with just talking about the different tools. So we’re gonna be testing. So we’re gonna be testing five different tools, we’re gonna start, we have outranking.io, which they were kind enough to extend us a an account, go charlie.ai, which, again, they were kind enough to extend us an account, we signed up for a free trial of lately, we have a paid account with open AI, which is the granddaddy of them all, in terms of language models. I tried signing up for a free trial with Jasper AI, but they require a credit card, they say it’s a free trial. And then the bottom, it says actually, we’re going to charge your card money and the charge will be dropped in a couple days. So I said no, we’re not going to do that. So they’re not in the consideration anymore. And instead, we’re going to use copy AI. So these are the five tools that we’re going to use today. Several of these have similar features, like a one click blog post kind of thing. And others have slightly different features. So let’s start with, let’s start with open AI, which is probably the one that people have heard the most about, it actually doesn’t have much of an interface because the company is sort of the back end for a lot of these other tools. A lot of these other tools integrate with it, but they do give you this playground to try this out. Their software works best with a prompt. So let’s go ahead and come up with a prompt. I’m going to do this write a six. Let’s do right six paragraphs about the importance of social media ROI. Now, before I hit go on this, Katie, what are you expecting it to write about?
Katie Robbert 9:07
I’m expecting it to write something along the lines of what is social media ROI, why you need to consider social media ROI, what is ROI, which is return on investment, and maybe a little bit about how to calculate it. But the importance really being like, you know, maybe ticking down reasons like you know, so that you can set more accurate budgets so that you can resource so that you can, you know, fine tune your campaign so that you know what the heck happened. So I would expect to see those general ideas in there.
Christopher Penn 9:47
I think that makes good sense. I put together this outline. This is from a talk I did a couple of years ago that would essentially be kind of like, you know, what I would expect of a profession. Shouldn’t human what they should write about? So exactly what you talked about Katie, what is social media, right? Defining social media defining ROI? Why do we care about this? You know, how do you calculate social media ROI? And then what do you do next. And you may see a blog post like this on the Trust Insights website, possibly more than one, but this is sort of my checklist, my sanity check for what these tools are going to develop. If a tool is really well suited for building content that requires, you know, a minimum of editing, it should come up with at least some of this right? One would think, all right, so let’s go off to the races. We’re going to start with open AI, we’re going to let’s crank this up and give it as much room to run as it can and see what happens. Let’s see, number ways to measure track leads and track sales track number of new customers. Okay, so it did five paragraphs instead of six. But what it’s saying here is not entirely I mean, it’s not wrong, the starway social media is great way to justify your budget show the value of your activities. So I think that’s that seems reasonable, right?
Katie Robbert 11:14
I think it does. And so, as I was sort of saying earlier, in terms of my personal use case, this is the kind of information so I’ve been using open AI because that’s the tool that we have an account with. This gives me a general outline as a jumping point. So I would take this and probably rewrite 90% of it, but at least it gives me an outline and a structure to follow this logical, sort of breaking down that, where do I even start? There’s so many things I could say, this at least gives me that okay, here are five points that I can make about social media ROI, and maybe I can expand upon it in a different post.
Christopher Penn 11:54
Right. Okay. So that’s, that’s our first example. And that’s, you know, like I said, That’s a decent start. One thing I think, is really important to mention here, and we’re going to try and get them on one of our channels shows at some point, but I had a chance to chat with Attorney Ruth Carter of She’s based in Arizona, they’re based in Arizona, and they said, the law about AI generated content is reasonably settled. And what alarmed me was this purely AI generated content is in the public domain, because it was generated by a nonhuman. And they said the law in question is references the chimpanzee that took a selfie. This was a few years ago, a chimpanzee took a selfie. And the photographer filed a copyright and what was going to charge newspapers to run this, and it went to court and the court ruled, the photographer doesn’t own that photo, the photo is in the public domain, because a non human took the photo made the content, and therefore copyright does not apply to it. And so Ruth Carter’s explanation when I was chatting with him at Content Marketing World was, if you just use the output of an AI, just copy paste it, it is in the public domain, you cannot copyright it.
Katie Robbert 13:15
I would be I would be surprised if anybody took this exactly as it is, without editing it at all. So once you start to edit it out even a little bit, I would assume that that then nullifies that the machine wrote it, it does not. What if I edit 90% of this
Christopher Penn 13:34
is if you had 9%, than the 90% that you edit it is is then yours, what you edit what you as a human touch is yours, what, but if you leave if you left, like the last paragraph, you end up with this, essentially a differential copyright where the first four paragraphs are owned by Trust Insights. And the last paragraph is public domain. Now, the caveat here is someone has to prove that in court, right. But just from an intellectual property perspective, and again, I want to get Ruth on one of our shows to talk in detail about the law and how it applies. Because with all the things happening with the image creation tools that raises really big important questions about what do you actually own, but that’s for another let’s literally for another show. So let’s try let’s now go into outranking. So, we’re gonna go into the AI first draft, our focus topic keyword is social media ROI. And let’s see for the advanced settings, the importance of social media, ROI, language, English locations, United States, let’s see what happens.
Katie Robbert 14:41
So, John, you know, I’m for the sake of this episode, I’m putting you in the same non technical bucket as me versus Chris’s technical bucket in terms of user what kind of usability or lack of information architecture in it interface is acceptable for your standards like, how far are you willing to go to learn how to use the system versus something you can literally just plug in a keyword? And then just go?
John Wall 15:13
Yeah, I mean, that would just be a matter of digging in and playing with it. I mean, if, you know, I can definitely go beyond just pasting stuff in if there’s parameters to adjust if it’s going to make a difference. Like it was interesting in that last piece that there was plenty of stuff about ROI. But it actually didn’t say anything about social media. I mean, you could have just switched it to SEO or direct mail. And it’s all still valid, you know, there’s nothing about social media actually in there. So that’s the big thing is, you know, how, how accurate and tight and useful is the content going to be? Or is it just going to be so generic that it’s not really, you know, saving you some typing time, but that’s about it.
Katie Robbert 15:50
And while this is generating, I would like to sort of do a little bit of a PSA, to any of the viewers of this show, please, please, please do not take this show as an opportunity to then slide into our DMS and pitch us your product, this is not the purpose of this. Obviously, there are there are tools out on the market, this is just a sampling. So please, I need it when I say do not start pitching us your tools.
Christopher Penn 16:17
This is 31%. So let’s move on to the next fuel wallets working. We’re gonna go to go, Charlie, now we’re gonna go, we’re gonna copy and paste the exact same title, which is the importance of social media ROI, our target audience, let’s put in chief marketing officers, and our SEO keyword will be social media ROI. Let’s see what go Charlie comes up with. We’ll get that going.
Katie Robbert 16:40
It’s this tool. And I haven’t dug into this one at all makes me do a little bit more of the work of coming up with the title in the audience. Whereas part of me was like, No, you tell me what that is. But it makes sense. Like, the more focused you are, the better quality content you’re likely to get back.
Christopher Penn 16:59
Exactly. And I forget how long this takes. So let’s pop into so lately is an interesting one lately doesn’t have the exact same interface. Instead, it’s going to try and create stuff. So I’m going to take an episode I did all on social media ROI and see what it does. So let’s take this, this is an episode I did. This is going to take a little while as well. And then let’s go into copy AI, and create our first blog post. So enter our blog post title here, let’s do the importance of social media ROI. Our key word are tone. What tone do we want? Katie?
Katie Robbert 17:37
I think professional probably makes the most sense.
Christopher Penn 17:40
Yes. Yeah, I was gonna go for wedding.
Katie Robbert 17:45
Well, if it’s supposed to be an educational piece, you know, we can add? Well, and then here’s the thing. That’s the kind of decision I would want to make for myself. Because my style of writing and the way that I insert humor into my content, I don’t want a machine to try to do that for me. I just want facts.
Christopher Penn 18:09
Yeah, that’s fair. Okay, so let’s come up with this section, the outline, let’s see about generating the talking points now. So this is sort of doing it more sequentially. And this is interesting, because it kind of it’s trying to, I think, generate, essentially, a more complex outline to keep it on the rails. Remember, we were talking about transformer models go off the rails pretty easily. So it’s going to come up with this outline. So what is return? What is ROI? Okay, good. How do you calculate it based on your goals? Why is it important? And then sort of a wrap up, paragraph, let’s go ahead and hit generate content. Alright, it looks like it’s gonna be going. Okay, so we have, we want to go back to outranking. Now, let’s see, let’s choose a title. Which title do we want? So many to choose from?
John Wall 19:02
That’s so funny that everyone has a free calculator. Like, you know, those calculator pages that are everywhere.
Christopher Penn 19:08
I guess let’s do the first one. That’s fine. Let’s see outlines from ranking pages. So this is essentially borrowing this to create this outline, what is it? How to measure it, when to measure it? What to measure? So forth. Okay, cool. Let’s congeneric SEO brief. Premium first draft. All right. With that, what that button does and says this was can take up to 15 minutes. Awesome. So we’ll move on that do its thing. Go Charlie has done its work. So let’s see what we have here. The importance of social media ROI even came up with a nice, you know, stock image here of anonymous people laughing at something. Sure. Exactly. See, is our introduction. What is it man? To return on investment in social media, it’s important to care about it, number ways to calculate it, how to choose your business goals, making a case for it, why social media should be a key part of your marketing strategy. And then what metrics to use. So this is, this section here is factually wrong. Social media ROI is a computation earned minus spent divided by spent so that section here has nothing to do with ROI. See how to maximize inclusion. So this post is definitely more detailed than what open AI came up with, I think would require a lot of editing. Because the language is fairly bland,
Katie Robbert 20:39
I would assume that they would all read, I mean, I would hope that they would all require editing because every single company, every single person has their own unique tone and voice. And that’s a big part of why I’ve read from certain people. You know, if I’m just getting this very, like, monotone version, I’m not gonna read it. I’ll be bored.
Christopher Penn 21:05
Yep, it looks like I can change the tone here as well. So formal formality, casual formality, increased length, cowboy, Rick V.
Katie Robbert 21:13
Let’s go with science, Rick. Okay, you want this to be useful? I guess? I don’t know.
Christopher Penn 21:23
I guess I let’s let’s have it regenerate with that. Sure. Lately is still thinking. Alright, so we have copy AI has come up with so here’s an introduction. What is it? What is ROI? Let’s see. This is correct. So this is earn my spent have ever spent, so they get a bonus point for that how you calculate it? This is somewhat correct. There’s some issues with with the math in there. But so okay, why is it important? It matters and then your conclusion. So let’s go ahead and create that blog post with that professional tone. And and here we have a first draft with confetti, no less,
Katie Robbert 22:09
you know, so it’s interesting that as you’re going through it saying that things aren’t factually correct. I, again, sort of like sort of the use case for me is, I see it as a jumping off point, I expect to have to edit things. So I’ve written I’ve used open AI to write a couple of blog posts so far. And the advice the quote, unquote, outline it was giving, I was like, Well, no, that’s not the advice I would give, or that’s not the way that this thing should be structured. But again, what I’m finding is having something to react to versus creating it cold is so much faster. Because it’s so I feel like as humans, it’s so much more efficient, to want to correct what we see is incorrect. So if you look at this and go, No, it’s all wrong. Let me do it right the first time you just write it up, it’s like, oh, that’s way less time. Because you’re like, let me fix the thing.
Christopher Penn 23:07
Ya know, and I know for some people, some people are better creators, and editors, and other people are better editors and creators. And so if you are in the second camp, this is an invaluable way to either get started or get over some writer’s block where you’re like, I don’t know what to write, it’s easy for you to have a machine, right and then go, No, that’s not what I write. And it sort of hijacks your brain into into saying, This is how I would do it. And in turn, you can create that content. Let’s, I’m gonna go ahead and take a look. Let’s see what happened here. If science Rick did anything different. I don’t really see anything different about using science works. So okay. We’ll go ahead and just store that store this. Now let’s see how outranking is. So outranking has a we have what is social media ROI? That’s some artifacts in the text. They’re setting up your metrics value of metric, that’s yeah, okay, that know how to measure it, when to measure it, how to build a better social media strategy in 2022. Okay,
Katie Robbert 24:18
well, this goes back to where are these systems, these programs getting their information? And I think that’s one of the you know, buyer beware tales. And I’m not saying that any one of these is better or worse than the other. But in terms of how much editing you’re willing to do and how much subject matter expertise you have on the thing you’re asking it to write those are considerations because if you are asking it to write about something that you don’t know all that well, that is a huge risk.
Christopher Penn 24:57
Yep. Okay, let’s see what we got here. So late. li Oh, lately a string summarization of of text content. So it took it took a social poll, it took a blog post in his trunk to social posts, so it’s not generating that new text. Okay. So that’s not going to be a good apples to apples comparison.
Katie Robbert 25:19
No, but it is a use case that I can see that, you know, some people would want to have it’s like the abstract of a research paper. What the heck is this thing about? Do I care about the topic before investing all of that time to watch this video or read the thing? So that is still a good use case? And it looks like you know, so I see the Twitter icon at the top, like, okay, it’s gonna give me like a little social post snippet. That’s a really helpful thing.
Christopher Penn 25:50
Yep, exactly. So that’s what we’ve come up with, we’ve got four completed, posts, four completed articles with very different styles, I think.
Katie Robbert 26:06
Well, in turn by the AI, not completed,
Christopher Penn 26:09
yes, completed. So we have four first drafts, let’s call it that way for machine generated first drafts. And now, it’s up to us to to take this and take it to the next step. Now in terms of speed, open AI is GPT-2, by far was the fastest, right we watch it generate in real time. And it was really, really fast. Charlie was the second fastest. I would say outranking, and copier AI seem to be about the same. Charlie’s had the least setup of the UI tools, right? There were more steps and copy AI and outranking for keyword and stuff. Whereas go to trial, there’s like, Okay, you just give me the keyword and I’ll come up with something. In terms of quality of output, like you said, they’re all good enough to start editing.
Katie Robbert 26:59
Well, and it’s funny, because with open AI, yes, it was the fastest, but you’ve got the least amount of content to work with. And so those are the trade offs of, do you want something that is super polished, and already, you know, has like so this one, this is outranking, it gave you like the h2 headers, and it gave you you know, all the way that it should look at the formatting. And so then you have to spend a little bit more time waiting for it to generate in order to get that. But the trade off is the work is mostly done. Now you just need to make sure that it’s accurate and in your own company’s brand and tone.
Christopher Penn 27:40
Exactly. And it does have some other neat stuff here. Like I just I hit optimize, and saying, hey, the surprise, some additional things you might want to add into this post.
John Wall 27:49
Yeah, they have a real SEO slant, don’t they? I mean, that’s their number one concern on this,
Katie Robbert 27:54
which makes sense. I mean, that at the end of the day, theoretically, that’s a big part of why you’re writing more content is to boost your SEO.
Christopher Penn 28:03
Yep. Let’s see if copy AI has something similar. So you’ve got tools, popular tools, Instagram captions, so that’s not helpful,
Katie Robbert 28:15
but not in this instance. Yeah, exactly. What is your like, as you’re watching this all play out? Like, how is this making you feel as a creator?
John Wall 28:28
Yeah, you know, it doesn’t hit as hard as creator. But as someone that’s doing any kind of curation. I mean, it’s huge, especially like that ability to summarize other posts and articles and just have it go right to social. I mean, you could literally just crank those out, and then you load those up into Hootsuite. And there’s your social summaries. And that’s real curated content. It’s not just like, taking sound bites. And then yeah, totally from an SEO, I hadn’t thought about that much. But this idea of just loading up, you know, the 20 best pages from H refs, and then turning that into your own content on your own site that’s fully optimized, that just seems like a no brainer that you could bang through.
Christopher Penn 29:15
So what are we going to do next day?
Katie Robbert 29:20
Well, what we’re going to do next is use these four different pieces of content to write four different blog posts and post them on our website. But in terms of if we were had to make a decision here, and now in terms of, I need a tool that’s going to assist me in writing even bigger volumes of content that’s high quality, I would start to look more seriously at a tool like outranking versus open AI, which is what we’re currently using, because it gets me farther along in the process. I mean, so go Charlie does a similar thing. And so those tools is those. That’s what I would start looking Yeah, in terms of making those decisions of, okay, I need to weigh the cost against the benefit. The cost might be, you know, however much a month, but the benefit is that I’m able to then crank out like an additional 50 pieces of content a month, and improve my awareness, and then my business can grow even faster. So that subscription fee might be, you know, minimal compared to what I’m able to bring back.
Christopher Penn 30:30
Yeah, I mean, having gone through this, now, we’ve gone through the exercise here, I would say, of the different tools, if, if I had a team of people who were fairly junior, I just needed to do a lot of first drafts and stuff, I would say, go, Charlie, the one click blog experience was super fast, and it generated good enough content, that then you know, a subject matter expert has to tune up, out ranking, I thought had probably the most thorough approach to generating really good content. I think it requires more subject matter expertise of stuff like SEO to begin with. So if you’ve got like really junior people, this might not be the tool for them. But if you have people who are proficient at content, and they just need something to get them going faster, get, you know, to a first draft faster, I think outranking really fits the bill there. Either one of those tools, I think, you know, in terms of helping you create content faster would be a good choice. I think open AI is the right tool, if you’re building your own software, right, if you’re gonna build something and hook straight into their API and stuff, the playground is fine. For for what it is. But like you said, if you if you want somebody who has less experience, being a content creator, either out ranking or go, Charlie would would be better choices than open AI.
Katie Robbert 31:57
See, it’s funny that you talk about it as experience, I look at it as time. So I think, again, open AI is a good tool. But in terms of time savings, I think outranking and Charlie AI are the things that for me, my personal use cases, would probably save me the most time, once I dig into them to see how much editing these tools these things need. And so that’s sort of where I would need to continue continue to do my own head to head have, you know, is open AI giving me just enough to work with that I can just write it and be done. Or am I spending all of my time rewriting what tools like outranking and Charlie AI are giving me and so that’s, you know, still that comes down to your personal writing style, your bandwidth, your expertise in any of these given topics. And then just really, ultimately, what your goal is. So if your goal is to write more content, look at your own process to see where you have the most deficits. And that should start to guide you into which one of these tools and how far along the process it should be able to take you.
Christopher Penn 33:06
Yeah, I would say though, that the thing that they’ve all not done well is the factually correct stuff, which, again, speaks to who’s doing the editing, like you obviously cannot and should not just copy paste straight to your blog, that would just be a terrible idea. But if you have somebody who is particular, if you’re working say like in in something that is not as as I guess, common a field, like maybe you were doing something on neurosurgery or you know, surface to air missile systems or something, you would need to have somebody with actual domain expertise, reading the content and saying, and that’s, that’s not great, because if you were, if I handed this to my 17 year old, they’d read it. And because they don’t know marketing, this again, that makes sense. And whereas we’re sitting here going, that’s completely the wrong answer for what ROI actually is. But you have an uneducated reader, or an educated content creator, maybe the most junior person on the team that you just pulled out of college. There could be issues creating and putting up content that would pass a human editor who didn’t have the domain expertise.
John Wall 34:17
How about over on lately? That was the one that did the summary. Right? Does that have? Does that have other options? Or is that really all that thing does?
Christopher Penn 34:26
Um, I think that’s it.
John Wall 34:29
Okay, so that’s really that’s a good curation tool. But that’s not going to solve your write a blog post problem,
Christopher Penn 34:35
right? That’s correct.
Katie Robbert 34:38
But it definitely has a strong use case because once again, sort of I’m not a social media person. Having a little bit of a leg up on something to react to is going to save me time when I’m trying to write the blurbs now, am I looking to use like four different tools? No, because that actually would probably take me longer than actually just sitting down and doing the work my Self? You know, and that’s part of it is like, are you procrastinating? Are you looking for these tools to help you stop procrastinating? Maybe. But I don’t think you need to, that’s a different problem, the AI isn’t gonna solve that. That’s a people problem.
Christopher Penn 35:16
Yeah. Now the other thing about the open API interface, and the model is that it has a bunch of different functions that you have to read the documentation to deal with. But it has more flexibility. So I’m going to type in TLDR, here saying too long, didn’t read. And it will now take one of those posts we just wrote, and just summarize, right, so it will kind of do what lately just did, which is come up with a fast snippet. If I shrink, shrink that maximum length down to something along the lines, macing, like 1000 characters at most, let’s go ahead and tell it regenerate. And so now, it’s, you know, it can take and sort of distilled down content. So if it’s as long as it fits within, its full. So again, it’s got a lot of flexibility. But it is definitely not intuitive or obvious. The other tools are definitely much more intuitive and obvious, and how you can see that, you know, the design for all the different use cases available to.
Katie Robbert 36:17
And I think, again, that comes down to, you know, the trade offs of Do you want to spend. So and this is where Chris, you and I would approach it very differently. Your inclination is to move towards something like open AI, because you’re willing to read all the documentation and have more control over the features, whereas I’m not that person. So I’m going to look for something that has more of an easy to use interface that the buttons tell me what it’s doing. I don’t have to then go read developer documentation to learn how to write prompts.
Christopher Penn 36:51
Exactly. So at the end of the day, I think it comes down to taking things for test drive, doing requirements analysis, and saying like, what, what do we want the tool to to do, and like with every other SEO and marketing tool out there, the requirements are gonna differ from organization to organization, person to person.
Katie Robbert 37:14
Well, and I think that that’s the big takeaway is do the requirements gathering, you know, if you want to use the five P’s to do that, that’s a great way to start. Because the way that I need to use the tool is likely different from how John needs to use the tool, or how Chris, he’s used the tool. John, you may only care about, you know, content summarization. Whereas I’m the one who’s tasked with writing the content in the first place. And then Chris is the one who’s tasked with maintaining these systems or understanding all the features. And so those are three different use cases, for the same tool.
Christopher Penn 37:49
Yep, so to wrap up, I don’t think we have a clear winner. But we definitely have two tools that for the non technical user outranking and go Charlie seem to be pretty, pretty decent. We have lately as a summarization tool, open AI as the the Swiss Army knife with no manual as being where we’re gonna go. If I’m gonna rephrase what Katie said at the beginning of the episode, if you want to pitch us your tool, we are accepting sponsorships, or the Trust Insights newsletter and our podcast for not pitch me. Yes, contact John. Arranged sponsorships.
John Wall 38:31
You can have an episode next week this can happen.
Christopher Penn 38:36
And also full disclosure. outranking and goto. They both did provide free subscriptions to us. So from an FTC disclosure perspective, they did that the other ones we either pay for or that did not any final parting words.
Katie Robbert 38:52
You know, I would say that, from what I’ve seen so far, these tools are incredibly useful. I plan to continue to test them and use them however, I do, sort of do the buyer beware, make sure you understand the topic you’re writing about, so that you can check for accuracies and make sure that you’re planning on editing the content, for your own tone your own brand. You will get in big, big trouble, just in general with like Google algorithms and your readers if you start just copying and pasting exactly what the tool spit out. But I do think that that first draft, it’s those are pretty good.
John Wall 39:33
Yeah, these are tools for experts. You know, they’re not expert tools, that’s for sure. So go out there and be careful.
Christopher Penn 39:40
Exactly. And you know, machines not gonna take your job anytime real soon. But today is good. Not today. But it is good for all of us to know what is possible. And what is what is coming down the road. Thanks for tuning in. Folks. We will see you all next week. Thanks for watching. Check today. Be sure to subscribe to our show wherever you’re watching it. For more resources and to learn more, check out the Trust Insights podcast at trust insights.ai/t AI podcasts, and a weekly email newsletter at trust insights.ai/newsletter Got questions about what you saw on today’s episode? Join our free analytics for markers slack group at trust insights.ai/analytics for marketers, see you next time.
Transcribed by https://otter.ai
[postsignaure]
Newsletter
Podcast
