DocsBot: Automating WordPress Support with AI
Download MP3It's the WP Minute Plus, your home for long form discussions around the WordPress ecosystem. Deeper insights and conversations that you won't find on that other five minute WordPress podcast that we do. Follow both podcasts at the wpminute.com/podcasts. Big exciting news at the wpminute. You can now join our Slack membership for as little as $5 virtual coffee donation.
Speaker 1:Head to the wpminute.com/support. Purchase your spot be in the new refreshing Slack channel for WordPress professionals. Our annual membership still exists if you want to find some great annual member perks, but if $79 a year is a bit outside of your reach, you can now join that membership and support our work and get access to our great group for as little as $5. So what are you waiting for? Join your new home for WordPress professionals talking the news, business, and all things WordPress at the wpminute.com/support.
Speaker 1:The wpminute.com/support. Hey, Aaron. Welcome to the program.
Speaker 2:Thanks for having me.
Speaker 1:I don't wanna make this my AI therapy session. Mhmm. But I figured since I have you on the call, I might turn that into I might turn this into an AI therapy session. Aaron Edwards, docsbot.ai, imagine ai. But you have a storied background in WordPress.
Speaker 1:For those that don't know, catch us up to speed. Where did you start in WordPress? What have you been doing for the last decade and what led you up to what you're doing now?
Speaker 2:Yeah. Sure. I started let's see, way back WordPress 02/2006, I think, is when I first kinda joined the community. And I found it because of multi site. It was called WordPress Multi User back then.
Speaker 2:And I was building kind of a network of sites as a project. And so I started playing with that and then started teaching myself plug in development. And eventually, I joined up with WPMU Dev, which is a larger WordPress like plug in and services company. And I've been with them for fourteen years and the last eleven or so serving as the CTO for that company. And more recently, I've been kind of experimenting and building my own things, especially with AI technology and just went full time doing that in the last few months.
Speaker 2:So now it's full time working on DocsBot and Imagine AI, so
Speaker 1:it's
Speaker 2:exciting.
Speaker 1:I don't wanna go sidebar too much, but I saw recently they launched a sort of like white label reseller plan for for their hosting package. Were you were you a part of that in building out that technology?
Speaker 2:Yeah. For sure. We've been doing some amazing stuff in the last few years just focusing on targeting agencies and and products for WordPress agencies specifically. And so now you can do everything. All of the WPMU dev products can be fully white labeled and sold under your own brand from your own old own website and everything.
Speaker 2:It's pretty cool.
Speaker 1:Yeah. It looks pretty cool. Maybe I'll get somebody. Maybe you know somebody on the team. Yeah.
Speaker 1:I could get on the podcast. How about how about that?
Speaker 2:Sure.
Speaker 1:Let's talk about DocSpot. Let's start with DocSpot. Docspot.ai. I don't wanna pitch it for you. I'll let you pitch it.
Speaker 1:But I will say that I have found myself struggling to really get into AI, ChatGPT specifically. Can I find a use case for this ChatGPT thing that I pay $20 a month for? Can I really make it part of my day to day productivity? Because my god, Twitter tells me I should. Mhmm.
Speaker 1:DocSpot AI, what does it do? What's the value prop and how can folks get a hold of it?
Speaker 2:Yeah, sure. So after ChatGPT came out and some of these APIs first started becoming available, I started playing around with that and learning it a lot more in-depth. The And first thing I built was called ChatWP, and that's wpdocs.chat. Basically, I just trained a chatbot with all the WordPress core documentation, like from wordpress.org, and scraped all that and used that to train the chatbot just to make like a free kind of tool where people in the community can like ask questions about WordPress and get answers from its knowledge from that training. That was kind of like a fun project for the community and kind of a proof of concept also.
Speaker 2:And then I kind of took that technology and adapted it and kind of honed it further to actually launch a business out of it called docspot.ai. What DocSpot basically does is it lets you custom train ChatGPT with your company's data. Whether it's document files, whether it's just letting it crawl your entire website or your knowledge base or support content and using that to train the chatbot, which you can then turn around and use for frontline customer support and presales. Or you can use it for internal knowledge bots. A lot of our customers use it for that.
Speaker 2:For teams, they'll have their HR and standard operating procedures and things like that in there. And you can use it for, of course, copywriting and things like that. But I think the best use cases are for more question answer type thing.
Speaker 1:Is the secret sauce in all of this stuff like the actual training process? Like when you say I've trained it to do this, is that the secret sauce in into into how folks are going to leverage ChatGPT in any in? I want to make ChatGPT write better in the way that I write. Do I just keep training it on how I've written in the past for the last fifteen years online? Is that the secret sauce to make this stuff more effective?
Speaker 2:Yeah. There's a lot of different techniques for chat gpt. It was kind of revolutionary when it came out because actually the underlying model for that, which is they call it GPT 3.5, that's actually been around for quite a while for a few years now. No one found it that revolutionary until they did this new thing which was basically fine tuning it, what's called RHLF, and basically they fine tuned it, questions and then how humans would want to see an answer. And so they took that underlying model and they tuned it in a way so that it could be released in a chat interface where a normal person could ask it to do things without having to understand the weird prompts and special patterns you need to do to make it output what you want.
Speaker 2:That model actually isn't new, it's just the fine tuning that made it much more useful. And so you can actually extend that core public ChatGPT, you can extend that even further by fine tuning it We're using techniques like we do called retrieval augmented generation to where you're actually pasting text for it to answer questions based on. So doing that allows you to fine tune the output further and make it a suit your needs a lot better.
Speaker 1:Does OpenAI or ChatGPT, do they give you the I'm not familiar with this, I'll just say, do they give you the framework in order to help train these models on your own or are they like, no, no, you kind of got to feed that back into the chat GPT machine if you will?
Speaker 2:Yeah, well there's different ways. There's one that's called fine tuning and that was actually only released by OpenAI just gosh, I think less than a month ago or maybe about a month ago. And that is where you're actually like a fish training and creating a custom model. And that is better at not necessarily training with data, but more giving examples of styles. So it's kind of similar to if you said, you gave it a prompt and said, okay I want you to write this, here's a whole list of examples of what it should look like, you know, question and answer, and then you tell it to go to the races.
Speaker 2:So it's kind of like that, but a more efficient way of doing that so you fine tune it ahead of time with a database of examples of what you want the output to look like for a specific prompt. But that is not what we do at DocSpot. First of all, it's kind of costly and slow, and second, it's much better at capturing just the styles of output, maybe your speaking style or tone or things like that, than it is at capturing raw data and facts, which is what you want when you're doing a support chatbot. You want to be able to give accurate answers to your users. And so what we do is called Trevo Augmented Generation.
Speaker 2:Imagine if, let's say, you asked a question or a customer asked a question, Okay, what is the most expensive plan for this In the back end, what we're doing is we're actually searching all the documentation we've indexed and then finding text chunks that might answer that question. This is a text chunk that talks about their pricing. So basically all we're doing is we're pasting that into ChatGPT and then saying based on this text that we pasted here, what is the answer to this question? So it allows it to give a lot more factual answers without having to custom train it every single time. Does that make
Speaker 1:So it certainly does. So WordPress product company, they've got their this would go on and assist their knowledge base or their pre sales documentation. Maybe if they don't even have that, you could start like if it was a very, I almost said immature, but that's not the right word, early stage WordPress product company, maybe they just have blog posts.
Speaker 2:Right. Can
Speaker 1:we train it on on the basics of that and and that's how somebody could start?
Speaker 2:Yeah, that's the normal use case. Most people, they just point at their site and then we crawl the entire site map and index their blog and knowledge base if they have one and things are marketing information and that's usually definitely enough at least for handling the pre sales and questions about the product from your marketing material. And of course if you have an in-depth knowledge base that's much better if they're going to be asking very detailed questions about usage and things like that. But what we do is we make it really simple to have that as your frontline agent that's available 20 fourseven and can answer questions in any language that's thrown at it, is pretty amazing in itself. And then if it can't answer, it presents them an option to seamlessly transfer that to your existing live chat or ticket support so they can instantly of relate that to human support.
Speaker 2:The stats that we see are really amazing in it, like the companies that have implemented it well are seeing like about 80% of deflection rate. So that means eight out of 10 support tickets don't even hit your inbox anymore, which is pretty amazing and that saves a lot of time and a lot of cost for your company.
Speaker 1:Yeah. I can imagine anyone with a freemium model that has a free plugin on wordpress.org and you just get those tire kickers
Speaker 2:or bugs
Speaker 1:that just like no idea what they're doing, you know, this is something that that really helps. I'm on docspot.ai right now. I'm looking at the pricing page. Help me understand the DocsBot line item. So I see one DocsBot, three DocsBot, 10 DocsBot.
Speaker 1:What is the differentiator? What is a DocsBot?
Speaker 2:Yeah, essentially we just allowed you I wasn't sure if I should call it like a bot or a library or it's like kind of a combo of both of them right now. So basically it's an independent bot that's trained with an independent set of content and documentation. Then with each of those, you can configure different preferences like giving it custom prompts to change its behavior or tone of voice or changing the design of the chat widget if you're using a chat widget in your website, that kind of thing. Basically most businesses they just have one or two or maybe if they have like a bunch of distinct products then it's better for them to create a distinct bot for each product.
Speaker 1:What does the training look like from your customer's side? Is this something that they Okay. This is with Descript, right? Which is my audio editing app. I have to kind of train it my voice if I want to use the audio repair tool.
Speaker 1:And sometimes the audio training can be hours long, like you want me to talk for hours. What is it like to train a doc bot?
Speaker 2:Well it's really quite fast. So we have 15 different sources of data right now. So we have a single URL, you just paste in the URL and that takes five seconds for it to learn from all the data on that page. Document files, you can just upload text files like a Word doc, PDF, HTML, or you can upload just like a zip file with a whole directory full of your documentation. We have a feature where you can upload the export file from your WordPress site, the XML export, and that will just do your whole WordPress site like that.
Speaker 2:Our most popular ones are probably the sitemap and URL list, so you just point it at the URL of your site or a specific sitemap on your site, like if you have a sitemap, Yoast creates a sitemap for your documentation post type or whatever it is. So you can just point it at that and then our crawler goes out, crawls all the pages it finds on the sitemap, does all the fancy stuff on the back end to clean up that data and make it real clean and pristine and learn from all the text content, the images and links and tables and things like that and preserve all that formatting for you.
Speaker 1:Do they use their own chat GPT key or is that your own?
Speaker 2:Currently, yes. We have you provide your own keys so that way you kind of have ownership of your data and and control and things like that. And that way we're not, like, just being, a API reseller necessarily. We're not charging you some some premium on top of the OpenAI credits. Instead, we're just charging you for a service and then you bring your own key.
Speaker 2:We're looking at maybe changing that model in the near future just to make it easier for beginners to get started and have as like an optional ability for people to do.
Speaker 1:Talk to me about, putting the product aside for a moment, let's just talk about ethics in AI and I am not anywhere near versed in like ethical law in artificial intelligence. Although, what a title that would make on a business card. I could probably start handing it out and people would probably believe me. I hear something like this and I'm like, wow. You could train competitor content for yourself.
Speaker 1:Right? I could have a bot learning, let's say Gravity Forms just sucking up all the information that Gravity Forms is posting out, knowledge based blog posts, marketing, sentiment, all this stuff. And I can turn over to my competitor bot one day and be like, so what's Gravity Forms up to? What's WS Forms up to? What's WP Forms up to?
Speaker 1:And I could kinda just get these insights into into my own pool of knowledge. Holding that aside, I see people can block ChatGPT from crawling their site on through robots. Txt, no pun intended. But they can block that and stop ChatGPT from crawling it. Mhmm.
Speaker 1:Are we gonna move into a world where this knowledge isn't as easy to access right now for people to start improving their models? Should people be thinking, Wow, maybe I shouldn't allow AI to crawl my site. What's your thought on, thoughts on that kind of access to your content, to your data? Because competitors could kind of flip the script if you will and and do this kind of knowledge work against your own content.
Speaker 2:Yeah. As far as like that crawling ability, I don't see it any different than what search engines have been doing since the beginning and people pay a lot of money and spend a lot of time trying to get search engines to crawl their stuff, you know. I would start preparing for the future when it's not search words, AI answers and where people are gonna be entering in Google a question, what is the best forms plug in for WordPress and you want yours to be the one that it recommends, right? You don't want it to be your competitors that it recommends. Heck yeah, I would be doing everything I can to let chat GBT and other AI models crawl and learn from my content.
Speaker 2:And that opens, of course, all kinds of weird doors. Like, you remember the the old, like, black hat days of SEO and keyword stuffing and all the crazy stuff people would do to to try to get their sites crawled and get to the top of the results. And there's gonna be a lot of that stuff as people are experimenting with this whole new world of techno can I insert like secret prompts that that will trick the AI in the future to always recommend my product? Things like that, prompt hacking, kinds of stuff. Who knows what this new world is going to unlock?
Speaker 2:But in general, I think it's just the next phase of of what we've been doing before with SEO.
Speaker 1:Yeah. Yeah. It's it's it's quite interesting, you know, looking at, you know, ChatGPT four or may not all of ChatGPT, right? All of ChatGPT now can search the web through Bing's browse or Bing's search engine. Right?
Speaker 1:That was the the most recent update.
Speaker 2:And Yeah.
Speaker 1:Yeah. Just playing with that and seeing how you it'll cite the the source URL, you know, you're kind of sitting here going, geez, do I wanna put this content out there and and have this hard earned content be the source of somebody else's answers or somebody else's service? And then it's also, well, that door is pretty big. If it's search, if that's the future of search, now you're kind of kicking yourself if you lock yourself out. Right.
Speaker 1:It's a Wild West time to a degree in in my opinion.
Speaker 2:For sure. And I think kind of another thing too is it very much changes like how if we're like for us that are product owners and thinking about marketing and SEO is it changes kind of that strategy because we don't know for sure. Like, right now, like, Bing and and the Google bard, they kind of cite the sources. If they pull info live from from your website or something, they cite that as a source, so you still have that source of clicks. But, of course, there's gonna be other things and and general data is not gonna cite sources and and and it's just gonna provide an answer directly to a user and they have no reason to go and read your well thought out blog blog post, you know, about that topic.
Speaker 2:Right. And so I think that kinda changes the game for us as marketers in that our focus needs to be less on, like, informational things and probably more on things that drive intent, like maybe tools or or things where they would have to click through to take an action on your site rather than just pure pure knowledge and pure content marketing. Yeah. So so providing value in different ways to that, to that the AI can't do itself.
Speaker 1:Right. Do you have an opinion on on where the market's headed with AI rive to content. In other words, I don't know. I saw the other day some tool somebody posted that would create you a 100 blog posts and send them directly to your WordPress website. Right.
Speaker 1:Right. I'm just looking at that going, what the hell are we doing? Like how is anyone even going to, you know, you know, how how? Why? How do we get into this mess?
Speaker 1:So do you have an opinion or maybe some technical knowledge on what search engines are doing to safeguard search results from its own AI driven mechanisms?
Speaker 2:Well I think we have kind of seen with the latest like Google helpful content update how they have put in filters that maybe we don't understand but seem to be doing a decent job at filtering out, like, a lot of these content farm type things and even better. So, like, the thing has always been create helpful content. Right? Create content that people wanna read and that provides value to users. And as long as you're doing that for now, I think good, like, the SEO side.
Speaker 2:But as I said, as as people can get that stuff from the general knowledge of AI models that's being trained on that stuff, it's gonna be less and less of a way to get people to your site through creating that kind of content. And so, yeah, I mean, for now, like Google has said, they won't necessarily penalize something because it's AI written. Really, it doesn't matter who wrote it or what wrote it as long as it's something that's useful to users. So, I mean, I think that's what it comes down to. Like, you need to think about the quality of the content you create.
Speaker 2:It doesn't matter if it's AI helping you or not. And and keep that high and keep that useful instead of always thinking about trying to game SEO because that's always just gonna be a back and forth battle.
Speaker 1:This launch DocSpot launched six ish months ago?
Speaker 2:Yeah. Roughly?
Speaker 1:Yeah. And you're not you're not a novice to obviously the WordPress product space. Again, looking at the pricing page, 19 it ranges from 19 to $4.99 a month. $19 a month to $499 a month for the business plan. And then below that, we have an enterprise plan, $8.99 a month.
Speaker 1:I understand the world of of enterprise sales having worked at Pagely for a few years selling into the enterprise. How has this shaken out for you in terms of the the product packaging going into the WordPress space? Are there many people in the WordPress product space that are on that business and enterprise plan? How have how have how has this been received by, WordPress product owners?
Speaker 2:Yeah. I mean, most of my community and kind of network before this is WordPress and other WordPress product owners. That's kind of the first people that I was sharing this with and selling to. So we do have a lot of WordPress businesses that are using it. One of the biggest being automatic is using it for all their products, the happiness teams.
Speaker 2:And so, yeah, it's been interesting because this has been my first, like, true, like, b to b product, business to business. I feel like most and even, like, in WPMU Dev, all the products we built there, a lot of in WordPress, it's it's really, it's b to c because most of your customers are like, either they're just like, they have their own site or they're just like, maybe have a few clients and they're trying stuff out, you know? And so they can be very pricing sensitive. With like, I really didn't necessarily target WordPress specifically, just trying to go more traditional SaaS model. First launched the April was the enterprise plan and then I ended up having people that wanted more enterprising features.
Speaker 2:So I said, alright, let's make that business and then let's make higher on quote levels above that, you know, and that's worked out. I've I've signed a handful of really large ones like in fact the biggest one just started I think yesterday, that's Sony over in Japan. Congratulations. Yeah, which is a huge company and so it's been a big learning experience for sure learning how to deal with all the requirements and security and things like that that are required to sell to those large organizations. We try to have a plan that works for your regular WordPress product business size, and even one that works for personal use if you just want to do listening to your docs, question and answer with your own documentation or people use it if they're scholars or students or things like that, upload their textbook and be able to ask questions about it and stuff like that.
Speaker 2:Hopefully not generate answers.
Speaker 1:Man, I have three young boys. The oldest is in second grade so I can't even imagine a world of high school homework with AI.
Speaker 2:Seriously. I was looking at a photo of someone who used the new GPT-four vision, if you've seen that, and they took a picture of their textbook, math textbook with all the exercises in it and just stuck it in there and said solve this for me and it just created all the answers for them for all these graphs and everything in the math textbook. Was like, Oh boy, this is a whole new world.
Speaker 1:I'll know when my kids are playing video games in five minutes. Didn't you do your homework? Oh yeah, we already did it. I'm like, Okay, you used an AI bot today. You used DocsBot to do your homework, didn't you?
Speaker 1:Yeah. Quickly just going back to the enterprise and sales process that you've been uncovering, has the when you sell into the enterprise or you sell into the bigger business, not that you didn't have experience with this one before, but is there anything there that surprised you? Was it the legal side, the onboarding side, the procurement side of enterprise that was a surprise? Or was it I didn't realize this technical hurdle was gonna be here. Any surprises when you started selling into the enterprise?
Speaker 2:Yeah. I had some experience at WPME Dev. One of our big products is Campus Press. And so we have a huge amount of enterprise kinda like large and Ivy League and things like that. And of course they have huge bureaucracies and things at that level where they need all these security checkbox checked and things signed and all this stuff.
Speaker 2:It wasn't necessarily surprising. I haven't dealt with the sales side of that as before, it was mostly just the technical security side. So I felt prepared on that part, but yeah, definitely it's been a learning experience, just learning, okay, just the fact that you put an enterprise price up there, even though I'm not necessarily offering much of any added value is like, oh, yes, we're a big company, so we need to pay that. And they assume that automatically. So I'm like, okay, I'll just keep increasing price there, know, whatever number you put there is what they think they're gonna pay.
Speaker 1:It puts the whole like, you know, the plug in business, the the typical plug in business, you know, 59, 79, 99, right? And people are like, woah, 199, I can't believe anyone would ever buy that or pay that for a plug in, but then you get into higher ed. I mean, not only are they used to paying that kind of money, but I'll tell you something, if you ever try to sell into higher education, it takes a year, sometimes two years, just because of the cycle of how that organization buys. They have to get all the stakeholders, they actually literally have to wait for summertime to start implementing some of these products when it's not in the heightened school year, right? Purchasing processes for all these different verticals are outrageous.
Speaker 2:Yeah, what you're really charging for is all the bureaucracy, all the stuff that they make you go through.
Speaker 1:Yeah.
Speaker 2:because it's such a hassle compared to self serve.
Speaker 1:Yeah. And a lot of people don't realize that, you know, selling into big business, I mean you were, you probably, it starts out like this. Somebody at, you know, I'll just use you as an example, somebody at Sony, super hyped to use this, whatever, they're like IT manager lead somewhere in part of their organization, really want to use this and they talk to you and they're like, We're going buy it. And you're like, Great. Let's talk tomorrow.
Speaker 1:And they go, Okay. Procurement officer slides in like a Jedi warrior comes out of nowhere. Right? Then they start saying, Okay, let's talk money, let's talk legal, let's talk security, and then all these people come out of the woodwork and you're like, What happened to that guy over there that said he wanted to buy it? Where did he go?
Speaker 1:You know, it's a crazy world.
Speaker 2:Yeah, I had to fill out like a 10 page spreadsheet in Japanese about the security and and stuff, you know.
Speaker 1:Any other WordPress companies of note that you can talk about that are using this?
Speaker 2:Yeah, let's see. I don't know who I'm allowed to call out. On website, I had Ollie. And Ollie is a WordPress hosting and they were one of our first customers and they're using it for their frontline customer support. I know Extendify is using it, I think within their plugin actually.
Speaker 2:So that's kind of an interesting case. You can put our widget on your website, but we have really powerful APIs so you can integrate it directly into your WordPress plugin dashboard or however you want to do it. A lot of different ways to go about it. I know Automattic because they have a different happiness team for GluCommerce and one for wordpress.com. Some of them are using the chat widget because that's simple to get started and you can customize the design.
Speaker 2:Some of them are using our API in a much more custom way that is transparently integrated into their support workflow. And some people are using it in their existing inbox. So for example, I have it hooked up to our Help Scout, which is our support ticket software, and it automatically writes a draft reply whenever we get a request in there. So you can do things like that. I believe Sony is using it for an internal CRM tool to provide automatic replies and stuff, and that's fully just using the APIs for their internal use case to write replies to customers' communications and things like that.
Speaker 2:Very flexible. It makes it a challenge to figure out the best way to market it because of all the different ways it can be used and trying to figure out, okay, who is our ideal customer and how do we want to present this? Once you have I mean Because everyone knows ChatGPT, can do so much with it, not just writing content. So it's kind of like once you have that custom trained on your information, it's like, what can't you use it for? We're just still discovering use cases today.
Speaker 1:And I also see in the enterprise plan you have some Azure OpenAI service. I'm a fan of Claude AI. Use that quite often for writing. Find it to be a little bit more creative with writing versus chat GPT. Yeah.
Speaker 1:Plus I just like the fact that I because obviously I do a lot with podcasting, so I do a lot with transcripts.
Speaker 2:Yes. And you can paste
Speaker 1:Yeah, you can paste it. And I just something about pasting it in and it just turns it into a file icon. I don't know. Something about that just makes me feel like, okay, you got this Claude. In ChatGPT, it just struggles with a thirty minute plus transcript.
Speaker 1:But anyway, the point is are you gonna support other chat or other chat services?
Speaker 2:The plan is to I mean, the space is moving so rapidly. Right now, we have a lot of like OpenAI specific things like their function calling and stuff that we rely pretty heavily on that aren't available on other models. But we're definitely going to be looking into that. And hopefully there'll be open models that are much cheaper and accessible and almost as powerful. Know of the Facebook's Lama two, it's nearly at the base ChatGPT 3.5 level.
Speaker 2:The reason we haven't done that is most of our customers are international and a lot of those models are English only. We're not very good at other languages so that's why we're another reason why we're heavily OpenAI still.
Speaker 1:One last bonus question here. One of the questions I actually tweeted at Matt which he didn't answer, but I know Jetpack, I put a video out on the YouTube channel about Jetpack's AI features, which when I spoke to a rep at WorkCamp US, they said it was powered by ChatGPT. I'm curious why I don't see more, or maybe you have an answer to this. Why not some of the LLMs from Hugging Face? Which, if I have this right, is more open source friendly than OpenAI.
Speaker 1:I'd assume like maybe Automatic would want to align with that kind of open source methodology versus an open AI, but maybe I'm not well versed in it.
Speaker 2:Well right now it comes down to quality and speed and cost, right? Probably more than anything. The open models are not quite at the quality level that you would want for a lot of product. It might work for simple chatbots, but it's not good at following, like, as well. Or for example, like the Jetpack, I know, for example, they have AI in their form in the Jetpack forms, and so you can type in a prompt, okay, create a form that does this, and then it creates this whole form.
Speaker 2:And that on the back end is definitely using OpenAI's function calling, which is basically a way to get structured API output from human text input. So that doesn't really exist in other models. You can fake it by doing some fancy prompt work, but it's very error prone because it's not fine tuned for that. But the space is changing really fast. Every month there's a new model, a new model version.
Speaker 2:Mean a year from now, the landscape is going be completely different. It's a challenge to keep up with.
Speaker 1:Yeah, can imagine. Docspot..ai. Aaron, where else do you want folks to go to check this out or say thanks for doing the podcast?
Speaker 2:Yeah. You can always catch me on Twitter. It's ugly robot dev is my username. And we also have kind of some fun AI stuff that that I do called Imagine AI. That's mimajinn.ai.
Speaker 2:And we have that's like imaging technology. So we have a WordPress plugin where you can create AI images right within WordPress. And we also have a really fun, like, children's book generated with AI, where you, like, upload some pictures of your child, and it actually illustrates them into the children's books. You get a custom hardcover children's book in the mail. So that's a fun thing my wife and I made together and we're going to be pushing that for the holidays, so keep that in mind if you need a gift.
Speaker 1:Awesome. And that's going be at the website.
Speaker 2:Yeah. Imagine AI. Sweet.
Speaker 1:Yeah. Awesome stuff. Everyone else, thanks for listening today. It's the WP Minute Plus. Subscribe to the newsletter.
Speaker 1:Thewpminute.com/subscribe. Search for us on your favorite podcast app if this is the first time you're listening to this, and maybe, just maybe, I'll get more into AI in the future. Don't want it to replace my voice. I've tried those podcasting services, still not up to snuff. I have a job for at least the next six months, basically.
Speaker 2:Okay. What about translation? Have you tried the translating ones?
Speaker 1:You know, it's been a few months since I've looked at them. The last time I looked at them, they were really expensive. Mhmm. But yeah, I mean it would be wonderful to have, I know a handful of other services have been released where it will speak in a different language for you as well.
Speaker 2:Exactly, with the same voice. It's amazing.
Speaker 1:Yeah, with the same voice. And that looks interesting to me, but it's still like $350 an episode. Thinking, well, if I had more sponsors, maybe we can release a Spanish version. But until then,
Speaker 2:just me.
Speaker 1:Awesome stuff. Thanks everybody for listening. We'll see you the next episode. That's it for today's episode. If you're a WordPress product business and you want to get your Black Friday deal in front of our audience, including a YouTube video, head to our WordPress product Black Friday placement listing on the wpminute.com.
Speaker 1:Scroll to the bottom. Go to the footer. You'll see a little fire emoji next to the Black Friday deal listing. Go ahead and click that link, and you can purchase a spot in our classified listings, get a promo video on our YouTube channel, sponsor the podcast. See the Black Friday deal listing in the footer of the wpminute.com.
Speaker 1:Thanks for our to our pillar sponsors, Pressable and Bluehost. Thanks to our foundation plus sponsors, the WP World. Without these sponsors, support from our paying members, and you, the listener, the WP Minute wouldn't be possible. See you in the next episode.
