AI Transparency, Trust, and Ecommerce Impacts

Jonathan:
Welcome to another episode of Do the Woo. I’m your co-host, Jonathan Wold, and with me again today is your other co-host, Tammie Lister. Tammie, how are you?

Tammie:
I’m good, thank you. How are you?

Jonathan:
Apparently a lot cooler than you are. How’s the heat over there?

Tammie:
We are having the two days of English summer, which are the two days I wish we had air conditioning.

Jonathan:
Well, I hope you can stay cool over there. We have a great guest with us today. James LePage is joining us from a cooler part of the world. James, how are you?

James:
I’m doing well. Thank you for having me.

Jonathan:
Yeah, it’s fantastic to have you. James is the founder of WPAI, and we’re here today to talk about AI, eCommerce, WooCommerce, and the future and all those fun things. But before we get to any of that, James, you’ve been in this WordPress space for a while now. Tell us about yourself. When did you first come into the ecosystem?

James:
Yeah, I basically grew up in the WordPress ecosystem, and I got started as a young teen. I had just worked a kitchen job; I was a dishwasher, and that proved to me that there was no dishwashing in my future. I really wanted to do something other than that. So I tried to figure out what I could do that wouldn’t involve dishwashing, but would give me my little allowance money to go and have fun around town as a kid. Somehow, I stumbled into web design. I originally got started with page builders and worked for neighbors and my parents’ friends, and eventually, it just grew and grew into something where I was working on more complex sites. Then I started selling to local businesses in my town, and that grew into me learning how to actually code. I learned how to code and really started freelancing as a true freelancer, eventually forming an agency called Isotropic. We grew to eight developers and built some pretty big projects, a lot of WooCommerce projects specifically for B2B. I continued that through college, then got into startups, did a startup called Share Club, and now here I am with WPAI.

So I’ve really been in WordPress since I got started in anything digital or business-related. My whole career has been in WordPress, and it’s been really fun to grow up in the industry and have my skills grow alongside me. It’s also been really fun to insert myself more into the community over the past couple of years as well because when I was running the agency and building the sites, I was definitely on the outside looking in. Now, running WPAI and building products for WordPress creators and users, I’ve had the opportunity to meet the people behind WordPress and WooCommerce and the communities around WordCamps and things like that. So it’s been really fun. Long story short, this has been my whole life and probably will continue to be.

Jonathan:
You have a journey that I think a lot of listeners can resonate with, where it started based on interest, and you built for others, especially if people started in the freelancer-to-agency space. You’ve been focused on product for quite some time now. Can you recall anything about that decision to switch from a focus on service to product? What motivated that? What was at the heart of that desire to make products?

James:
Well, the long story short is we all love building things, but I know a lot of agency owners and creators who don’t always love the client management side of it. That was kind of the thing for me. I love building sites, and I love creating projects. I come from a family of architects, and I think my whole family just loves creating. When you put clients into the mix, sometimes you can have the most incredible clients ever, but sometimes you can have really difficult clients to work with. We found that to be the case at Isotropic, where there were some really big projects that were pretty mission-critical to some larger companies, and it was just very difficult and stressful. So I started looking at products and how we could move away from that.

I also had a break in between my agency days and the WPAI days where I was a CTO at another venture-backed startup. So I was able to take a step back and really choose: Do I want to go into product, or do I want to go back into the agency and what we were doing? It was an easy decision for me to say, “Hey, let’s go try my hand at product.” I’d been using WordPress products for a long time. I’d been building WordPress companies and sites and things like that for a while, and it kind of just made sense to get into the product ecosystem. Specifically, for me, focusing on artificial intelligence was something I studied in university. So it was a really good mesh of everything. It was something that at the time, in 2022, was really just getting its legs and becoming the super exciting thing that it is today.

It was also something that I knew about, and I knew that by applying it to an industry like WordPress and WooCommerce—one that I really knew—I could really bring a lot of value by combining everything into one company. So it was a really easy choice to get into the product landscape. We launched our first product in late 2022, and it was CodeWP. It was something that a lot of people were very interested in, so there was also buy-in and traction from the community. It was like, “Oh, this is probably the easiest decision ever.” People really liked what we were working on; we were doing something that people seemingly wanted, and it wasn’t something where we’d have to deal with difficult clients. So it was a really good introduction to what grew into WPAI, and it worked out very well.

Tammie:
I’d love to start talking a little bit about AI specifically. There’s so much about the myth and concepts. If everyone’s describing AI, what does AI mean to you?

James:
That’s a very good question, and you’re completely right that there are so many different thoughts, ideas, and approaches to artificial intelligence. For us, we’re looking at AI as a way to help people do things that they’ve always done in WordPress and WooCommerce in a much easier and more efficient way. We’re leveraging a lot of the technologies that are being applied to many other industries and businesses, and we’re just using them to create tools that bring value to our industry. So for us, what AI means is a modern way, a new way to go and do business and enable the creators and users of WordPress in a way that couldn’t be done before. There’s the technical side of AI, and there’s a lot to unpack in that area as well. You could talk for days about how it actually works, how you can apply it, the different types of applications, and the various models and underlying structures of these models. But for us, our goal is to apply modern AI—primarily large language models based on Transformers architecture—to make our industry better and bring value to our industry. That’s always been our approach: sticking in the WordPress industry and doing what needs to be done for WordPress to make everybody’s lives a bit easier. That’s our approach and our focus on AI. But I mean, I’m happy to get into the technicals and all of it, but it might be an eight-hour-long podcast.

Tammie:
I guess I’d like to kind of build on that. I would love to go into technicals, and maybe we’ll get some opportunities in some later questions as well. But you mentioned ways to help; I’d love to also think about some of the limitations today. What are the limitations on being able to help, and the limitations on tools? Sometimes it feels like AI can solve everything the way that people speak about it or that it’s going to do everything. What’s something you’d like to debunk that people always get wrong about it?

James:
That’s a good one as well. I think that first off, AI is very hyped, and it’s rightfully hyped because I think it brings a lot of value to specific areas of industries. But it’s also not something that is going to replace everything, where you’re going to click a button and there’s going to be a brand-new website, or there will be a replacement of all these different jobs and things, at least in my opinion. Who knows what the far-term future is going to bring? But we’re in the early days of this technology, and where it really shines is in doing things that are repetitive tasks that can’t necessarily be automated away in a traditional workflow because it needs a little more leeway, a little more room to roam. But it’s not something that’s going to just completely replace everything.

In our approach with every one of our products, we’re creating tools that users—human beings—use to make their lives easier and to cut out the annoying stuff so they can focus on the fun stuff, like designing sites and actually building out the core structures and frameworks behind these WooCommerce websites and things like that, while cutting out the annoying tasks. For me, the reason we started CodeWP was that we had built WooCommerce sites for probably about five years at that point, and there were so many little extensions that you could make to the platform with these code snippets. We had this huge library of code snippets, and every time we wanted to extend a WooCommerce site, we’d have to go find the snippet, then go to the documentation reference, then spend eight hours on Stack Overflow, then go to three GitHub issues, and try to string together the information we needed to extend this snippet to extend WooCommerce.

That was not something that people really enjoyed doing. At least in our company, we really enjoyed thinking about how we could solve these problems. But

the difficult research and the challenge of stringing it all together were not something we enjoyed. So that was our first approach with AI: how can we cut out these tasks that we never really liked in our agency and make it so that’s something you don’t have to focus on anymore? Obviously, someone building a WooCommerce site isn’t going to use our tool to try to cut out the developer. It doesn’t work like that. You have to really know what you’re doing to leverage AI, but it’s one of those things where it does make our lives easier and allows us to focus on more fun things.

To get back to what people should be aware of regarding AI and what should be debunked here, it’s not a wonder machine. It’s not something that will do everything for you, read your mind, or anything like that. I think we’re very far from that right now. What it’s used for in practical applications is just to make people’s lives easier—those who are in the trenches actually doing things and getting things done. For us, the applications are pretty pointed because that’s where AI can thrive. If you try to tackle massive tasks that require a lot of logic, creativity, and thought, it doesn’t work that way. But to enable creators, builders, and WordPress users in our approach, you can really bring value without causing issues or stepping on any toes.

Jonathan:
You started out with this context of the agency and building WooCommerce stores, so you already had that sort of broad context. You’ve talked about repetitive tasks as a prime example of where AI can help. Would you expand on that within WooCommerce specifically? When you’ve had this service background building WooCommerce sites, and you’ve been in the product space for years now, so you have a broad context. If someone is thinking about AI within the WooCommerce context, is there anything beyond repetitive tasks, or within repetitive tasks? Can you just give us a broader perspective on where you see the opportunities for AI to be applied within WooCommerce?

James:
Yeah, so I’m actually really excited to be on this podcast. This is something that we’ve been really focused on with our upcoming product, AgentWP. We got started with CodeWP, which is an AI code generator. We looked at how people really wanted to use this tool, and there was a lot of usage towards WooCommerce. A lot of that usage was what we had essentially built it for—extending WooCommerce with code and cutting out the research bit for these really simple implementations, making it easier to build more complex implementations with code. But there were also a lot of requests outside of what CodeWP was intended for, in terms of creating product descriptions, giving statistics, troubleshooting, and all these different things. So a lot of the usage in CodeWP led us to determine that there’s a really big opportunity here for another product. This product could be one that lives in WordPress, lives in WooCommerce, and really integrates with that individual site, database, file system, frontend, and backend to give really valuable answers to users in a more generic way—something that’s not always code, but could be, “Hey, here’s your error log in English,” or “Here’s a statistic about your WooCommerce store,” or anything like that.

That research from CodeWP led to this product called AgentWP. That’s the product that goes into a WordPress site. We love WooCommerce, and we know how powerful it is, but it can also be pretty difficult to navigate sometimes and pretty complex, even as an agency, but definitely as an end user. So we’ve always, from the beginning of developing this AgentWP product, thought about how we could solve common WooCommerce problems. With this beta that we’re launching fairly soon, we came to the conclusion that WooCommerce can be extended in a few ways. Right now, the core way is statistics—really understanding your data. WooCommerce themselves have made really awesome improvements to the analytics side of things, in my opinion, transitioning from the scary Reports page to the beautiful React-based analytics page. But still, there’s so much data to sift through and even understand what it means. I think there’s a really big opportunity to have a natural language interface with your WooCommerce data and, zooming out, with all of your WordPress data as well, and leverage artificial intelligence to take a user’s request and turn it into a statistical response. That was a big focus of ours for the beta, to kind of prove out, “Okay, this product, the idea here is it lives in WordPress, leverages WordPress data, and then it uses or converts the user’s request into some form of action, response, or message in some form.” Our core implementation to solve this was with WooCommerce statistics, so users can ask anything about their store, and the agent will go and create a query on demand to get that information and then present it to the user. Then you can continue onwards and ask additional questions like, “Hey, let’s dive into this specific product and see if we can get analytics or insights about whether we should even focus on this product based on this data compared to all the other data.” These are things that you can’t really do with charts.

That was a really big use case that we wanted to solve for, and we did. It extends out to the remainder of WordPress, but we actually started with WooCommerce because that was something that in the agency landscape, when we were building WooCommerce stores, we always found issues with. And then in terms of building WooCommerce, we always would break it. We would always try to extend it, and we’d always install a plugin and have to do our plugin check to figure out which one was conflicting. So I think there are a lot of different implementations here, but the core ones for us were statistics and troubleshooting—not in truly the code sense, but just making it easier to find problems in WooCommerce because everybody who’s built a WooCommerce store knows that you could get very complex with these things. You could legitimately have a use case for 30 plugins sometimes, 20 plugins sometimes. In my opinion, that’s very different from standard WordPress development, where you try to keep your plugins down, but with WooCommerce, it just bubbles up. So having this interface where you can use natural language to navigate around your WooCommerce site and understand what potential issues could be popping up, that was something we really wanted to focus on. All of the development here was driven by the problems we bumped into while actually building WooCommerce for years and years.

Jonathan:
I want to expand on that a bit. So far, what I’m hearing is that there are two broad categories that you’re focused on. There’s the insights, which I think is a great example—being able to query against a large set of inputs and data to get insights for the end user. Then there’s also the repetitive tasks part. Those insights might lead to tasks to be done. I’m really curious about what I would describe as the opportunities categories. You alluded to it in terms of identifying potential problems and addressing those ahead of time, where the agent might be able to surface, “Hey, there’s an issue over here for you to know about.” But because WooCommerce is a subset, an ecosystem within WordPress, if someone’s in WooCommerce, there’s a lot more potential for knowing their intentions. They’re probably trying to run a business. There’s likely a commerce intention on the other side of it, whereas with WordPress, it’s a lot more generic. So I can see within the context of this product that you’re building, AgentWP, that you’re like, “Ah, it’s WooCommerce.” How do you think about the opportunities that you have—the ability to potentially surface to an end user—since there’s a lot more that you presumably could know about their intentions and the current state? Do you see that growing into things like conversion optimization or noticing, for instance, that customers are complaining about a particular thing and being able to… It’s broad, but there’s also a lot that you could know about intention and context because it’s in Woo. How do you think about that?

James:
Yeah, once you know that you’re in WooCommerce, there is a lot that you can really do for the user. Right now, the initial solution, the core idea here, was let’s just give AI safe access to the means and methods developers have always had. They can gather data, and then they can modify a WordPress or WooCommerce site using these APIs, queries, and all these different things. That was the core idea here. With the initial implementation of this agent, we started by saying, “Okay, let’s go and approach it from a read-only standpoint, where we can gather data and then use this data to tell users more about a site, store, or product.” That in itself is really valuable and beneficial, but our real goal here is to do both sides of the coin—do read and write—where you can surface opportunities and then the agent can take action on these opportunities for a user. In my definition, an AI agent is something that leverages its own internal models to do logic and reasoning for a user to fulfill a task autonomously. Obviously, you have to build rails and guidelines to make it safe, and you have to ensure the user actually wants these actions to happen. You also have to know how to identify this. But we want to grow this into that autonomous solution.

When you apply something like that—just the concept, not even thinking about the actual implementations—but the concept of this AI assistant within WooCommerce, specifically with access to this data and the ability to review the data for you, then the ability to give you suggestions and fulfill those suggestions, it becomes something that can be a real power multiplier for the agency building a site and the client running the site. Like you said, you can use the statistics to do synchronous reviews of your analytics, where somebody’s actually asking for a report or something and gets a report back. But you can also asynchronously review these statistics as well and identify, “Hey, here are some trends that

you aren’t even looking for that you should know about,” and “Here’s a product that’s not performing well.” By the way, we have this external data source that looks at the SEO for all of your products, and this one’s not really doing as well as all your other products in terms of rankings, keyword density, and all these different things. You can surface these opportunities in the read-only mode, but eventually, you can actually say, “Hey, can you just watch my products? And when a product starts falling in the Google rankings and when the Google Analytics metrics start dropping—the conversion analytics start dropping—can you take a look at this and see how you would improve it and then send me a message saying, ‘I’d improve it this way,’ or actually just go and improve it? Go and update that description to fulfill those keyword requests or whatever the SEO optimization would be, or rewrite it to be better for conversions or whatever it is. Go and actually do this for me.”

So that’s the end goal here, where you have this system that goes through your information autonomously and really understands the purpose of your site and how your site operates in terms of, “Here are the products that this company sells, and here are the plugins that this company is using to enable subscriptions and memberships,” and all these different things. Then really use that information to create those on-demand suggestions for the user and eventually notify the user about this. The user doesn’t even have to think about, “Oh, what should I even ask this thing to do?” Eventually, if the user wants—obviously you don’t want to just have this running around willy-nilly—but if the user wants and creates these rule sets, it can go and do these optimizations, and it can eventually go and do customer support or anything like it. When you give this system access to the data and the ability to take action on that data within WooCommerce and within WordPress in general, I think it becomes a very powerful and value-added system for both the builders and the users. It’s something that I’m really excited to be building because that’s what we’re working on right now, and it’s very fun to think about these problems and how you can build for our industry.

Tammie:
So we’ve spoken a little bit about the product and how you interact with it as a user, and we also typically like to ask how you would build here. I’d like to take a little bit of a tangent on that and say, how would you build integrating artificial intelligence into products? Because we spoke specifically about tasks and taking workflows. So we talked a little bit earlier about resetting that notion of what AI is and what it isn’t—debunking a little bit. What do you think it can solve today for product makers? How do you start using it, and how do you use it in the products that you make?

James:
I’ll mention a couple of implementations that make sense and how we would go about doing it. The first thing is I don’t think if you’re a product builder, you should just go and say, “Hey, we need AI. Let’s go figure out how we can add AI to our products and have a sparkly AI button in all our tools.” I don’t think that’s the case because AI, like we said in the beginning, doesn’t do everything, and it only does a few things very well. To do those few things very well, it has to be very mindfully implemented. You have to really think about how you’re going to, number one, gather the information necessary for these models to actually create something that offers the user value, instead of just garbage input, garbage output, or something that somebody sees but never actually uses. There has to be a very mindful implementation of the solution, and then there also has to be a very thoughtful way you actually implement it. So it has to offer the user value, and then you actually have to implement it so it does offer that user value.

In terms of products, I wouldn’t just completely pivot every product to say, “Oh, we must have AI now.” But I do think there are some really low-lift tasks for many plugins that can use AI in the statistics and in the generation capabilities. We actually see this with a lot of WordPress plugins nowadays. I just saw on X yesterday that Monster Insights added a Google Analytics AI chat where you can discuss your analytics. That solves a pain point. That’s something that takes data and makes it accessible to users, and it’s not this huge AI implementation where now we’re AI analytics and now we do everything, but we offer this helpful conversational interface to go and fulfill something that has been a pain point for users. GA4—nobody knows how to use this thing. It’s the most complex interface ever, so let’s go and simplify it. I think that’s a good implementation, and I think there are a lot of other products that have these pain points that could be solved with this open-ended conversational interaction with their product, but not actually going and saying, “We’re going to go make everything AI, and you’re going to have to use conversational interactions for everything. There’s not going to be a ‘New Product’ button in WooCommerce now; it’s going to be ‘Make Product with AI.’” That’s not the right way. But adding these little implementations, leveraging AI, and not going crazy and training your own models and building these data pipelines— which is something that we’re doing, but you don’t need to do that—you can use the generative AI capabilities off the shelf and build out these small implementations that solve user pain points that couldn’t necessarily be solved in another way.

I think another good example, not WooCommerce specific, is WSForm by Mark Westguard. It was one of the first form builders to implement anything AI-related, and it wasn’t this thing where every single input field and every single thing had this AI wizard and this “Generate with AI” button. But there was a very simple flow where, if you want to go and build out a form and you don’t know exactly what fields you want, but you know your concept—”I need this form to accept reservations for this specific restaurant”—you could type it into this natural language input box, and it would just convert that natural language into the underlying JSON structure of the form builder. Just a simple implementation, something that solves these problems, doesn’t replace the way the actual product works, but just augments it. I think there are many products in WordPress that can mindfully implement things like that, and that’s the approach I think every company should take with AI.

Just because it exists doesn’t mean you need it. There are many cases where it’s actually probably better not to have it, but there are also many cases where there are these small pain points like, “I need form inspiration,” or “I don’t know exactly what fields I should add,” or “I don’t know how to make sense of these big massive reports,” where a simple implementation that complements the existing product, instead of taking the product over, can offer people value. In terms of products, that’s the way to think about AI, especially with existing products. When it’s a new product, I think it’s a bit of a different approach, but with existing ones, you have the data, you know what your users want, and you can mindfully say, “Yeah, we should implement AI,” or “We shouldn’t implement AI.”

That’s one thing. The other thing is actually building the AI. With all of the solutions that I’ve talked about, off-the-shelf frontier large language models are usually all you need to add this feature, and they’re very easy to implement. But they become very difficult to implement when you try to solve these crazy, wicked use cases that are pushing the boundaries and something that somebody else hasn’t done before—something that’s not as simple as input-output. When you start getting very complex, then you have to start thinking about what else goes around the mindful implementation of AI. For us, with this agent product, we are trying to do a lot, and it’s kind of the complete opposite of what I just said with the mindful implementation of specific cases in AI for existing products. Our approach is to mindfully implement some AI capabilities with statistics and generations, and things like that.

But when you build a system that has a much bigger scope, there are many more edge cases, many more things that can go wrong, and there’s a lot more complexity in the underlying systems to actually enable these things. For example, navigation on WordPress, fulfilling tasks and actions for users. When you do that type of stuff, then you have to think about, “Okay, maybe we could technically build this, and maybe it’s wildly complex, but is this something that can be done safely 100% of the time? Can we catch errors before they’re introduced to a WordPress site? How do we actually steer the behavior of these models?” For us, it’s like we have to build out these humongous datasets and these massive databases, and then train the behavior into models and build out pipelines and things like that. So mindfully implementing AI and doing it in a safe, thoughtful, and effective manner, when you get complex, is another thing that you really have to consider. For the products that already exist, I really think thoughtful implementation of small AI features that are pretty simple but solve those pain points is the way to go, and from the tech side, it’s fairly easy. If you want to get crazy complex, you can probably build a really cool product out of it, but there are a lot more things that come with it in terms of being safe, being thoughtful, and doing it so your users are not negatively impacted by that implementation.

Tammie:
You’ve mentioned a couple of times the word “trust,” and I’d like to just dig into that. Sometimes AI can feel unexpected to people, and sometimes things can feel… How do you build trust into AI? Let’s take an existing product that doesn’t have AI and you want to start implementing it. You have a good reason to implement it. How do

you look at building trust with that user base? What is your advice?

James:
Wild transparency as the company. If you’re the company, you need the most transparent approach to this possible. You have to explain, “Here’s what we’re using your input to actually do, and here’s how we’re actually doing it, and here’s the additional information we’re taking from this site or this plugin, and here’s all of that data that’s going to this specific provider, and here’s our agreement with this specific provider, this data processor, and here’s how they use it. They don’t store it on their servers; we have this agreement. They’re not going to use it to train their big models.” Then when the data comes back, “Here’s how it’s used in your WordPress site, and here’s what you can expect from input to output, and there’s going to be no deviations because we’ve built these safe systems to ensure that this is what happens with this user input and with this data.”

So I think on the smallest scale and the biggest scale, being really upfront about how data is used, and then how the actual interaction with this AI will be used is really important. Also explaining—because for 99% of these implementations, it’s using a big AI provider like Google, OpenAI, Anthropic, or any one of these—explaining, “Here are the actual processors, and we have this enterprise agreement with OpenAI, so they’re not going to use data to train and they don’t retain this, and here are all of these policies.” I actually really like the… I don’t know if it’s true regulation around this, but the aspect of always being upfront with your data subprocessors, and actually taking that step further and explaining, “Here’s how the data is gathered, here’s how it’s used, here are the structures, and here are the pipelines.”

For us with the agent product, that’s something that when we launch, it’s going to be very clear: “Here’s the data that’s used. Here are the processors that are using this data, here’s what comes back, here’s what’s expected.” For us, it’s a bit more open-ended, so there’s a lot of user consent during interactions with our product—actually being asked, “Hey, we’re going to take this action. Do you want to take this action? Here’s what this action will do. Here’s how we got to actually creating this action.” For statistics, “Here’s the exact query we ran.” So if you want, you can go and get the data yourself and verify this, and give users the tools and essentially the transparency behind anything in AI. I think that’s the way you build trust. Also, build systems so you don’t make massive mistakes. There’s never an opportunity for a system to… I don’t think any system’s ever going to go haywire and say, “I don’t like this WooCommerce site; I’m going to completely rewrite it.” But I think there are cases where the AI can get misaligned, and it can go and take the wrong action, which could be destructive to a site. So building systems to catch some issue before it ever happens, having the user consent so they know what’s happening, know what’s gone into this action and what this action will actually do, and worst case, if a negative action actually happens, having rollbacks, having backups, having easy ways to undo something that was autonomously or synchronously taken by a system, that’s another really important thing. So I think transparency and then actually thinking about real users who are going to use this, so we have to build these systems that don’t just go and do it 95% accurately, but also cover the other 5%, which will come from AI, but make sure we handle it gracefully and properly to create a really good product that people do trust.

“I know AgentWP” or “I know this statistics solution,” or “I know this form builder does exactly this with exactly this data, with exactly these processors.” In the form of some input-output, “Here’s the source of that output,” and in the form of actions, “Here’s the action that’s going to be taken, here’s why it’s going to be taken, here’s what it will result in.” Worst case, we have robust rollback tools to go and undo something if it wasn’t what you really wanted. That’s the approach, I think, for any AI product. Just be upfront; don’t just add some AI bubble and don’t tell users what data you’re taking or start taking users’ data without their consent. Just be mindful and be an ethical product creator. I think that the trust should naturally come because you deserve that trust because you’re somebody that does it properly.

Jonathan:
Yeah, you put in the work for it. James, you’ve got an interesting background. You started out with the agency, you’ve also been in the venture-backed startup. With WPAI, you’ve also raised money, and you’re raising more. What I’m curious about is that the venture-backed startup was outside of the WordPress space. You came back into WordPress; you gave us some sort of clues already for that, but I’d love to hear your thoughts on what, for you as an entrepreneur and for other builders listening in the space, what’s some of the value that you see in the WordPress ecosystem? Because there’s a lot of risk with starting a business. There’s a lot that goes into it. You and the team have been putting a bunch of work into this. Where does your confidence come from that investing into the WordPress WooCommerce space makes sense? Why do you stay in this ecosystem?

James:
Yeah, I grew up in this ecosystem, so I know how big it is, and I know how far-reaching it is. When I speak with other founders, investors, and people in the startup world—and we truly are a startup, I think in the truest sense of the word startup, we’re an early-stage company that’s really pushing the boundaries in our industry with new technology—when I speak with other people in similar positions in different industries, they sometimes know what WordPress is. Sometimes they don’t even know what it is, and when they do know what it is, they don’t realize the extent of the industry. Being able to grow up in WordPress and realize, “Hey, we’ve done projects that are worth hundreds of thousands of dollars and have done hundreds of millions of dollars of sales through them,” you really understand from the ground floor how big the economic impact of WordPress and WooCommerce is. Then when you look at, “Oh, WordPress powers half the internet, and the nearest competitor is this massive multi-billion dollar public company that has only a few percentage points of the industry,” it’s really one of those industries that I don’t think many people realize how big it is. I think you kind of have to start from within WordPress to even understand how big it is. Once you understand how big it is, there’s so much opportunity to disrupt WordPress in a positive way and bring these modern technologies and solutions from the startup landscape into WordPress to really offer valuable products to the biggest cohort of internet creators in the entire world.

So for me as a founder, it was always very clear that this is an industry that’s really big. It’s always going to be underserved. It’s always going to be something that I think if you want to build a really disruptive and valuable product, you can, but you have to know the industry to be able to do that. You have to understand the pain points behind what goes into building sites and using WordPress. Just be from it. That’s what brought me back to WordPress because I saw and I had the opportunity to go really do whatever I wanted and build whatever product I wanted and serve whatever industry I wanted. I had the opportunity to say, “Hey, I think that this AI stuff will be very valuable in the near and long-term future. I can see in my brain exactly how this thing can be implemented on both a small scale but also a really big scale in WordPress, in WooCommerce, in these ecosystems, so I’m going to go and do it.” That was my approach there, and that’s why I got back into the WordPress industry. But I think it’s just really understated how big this industry is and how much capital moves around in WordPress and is connected to WordPress. When you look at the competition, it’s much more flashy and snazzy, and that’s why everybody goes down that route. But it’s not as big as WordPress, and it’s not as economically impactful as WordPress. There are true needs here, and if you can identify those needs and ideally build a valuable product that serves those needs—in our case, that’s trying to solve a lot of the pain points with AI specifically—I think you can build an unbelievably valuable company in this industry.

Jonathan:
James, if someone is looking to come into this space, maybe they see this opportunity, but they don’t have that background, what’s one piece of guidance that you’d give to an entrepreneur? Let’s say they want to build a WooCommerce-centric product. What guidance would you give them?

James:
If you’re from the industry or if you’re not from the industry, I think the only way to build any product—and it gets even more important with WooCommerce and WordPress—is to talk to the people who will buy the product. That’s something that a lot of founders don’t do, where they just have a cool idea and go build it and hope for the best. But specifically with WooCommerce, you can find a WooCommerce store, you can find the people who built that store, you can find the people who run that store, and you can just go to them and send them an email and say, “Hey, give me five minutes of your time, please. I really want to talk about

this potential problem that you very likely have that’s probably costing you a lot of time and a lot of gray hair, and I’m trying to build something to solve it. If I solve it, you’re going to benefit from it, so please talk to me.” People will speak with you, and they will spill all of their problems. Specifically with WooCommerce, I see because it’s a very scoped ecosystem, you could build anything, but it’s pretty scoped. People all have the same problems, and they all bump into the same problems doing the same things, and that’s where you can build valuable products—you can solve those problems. Those problems are surfaced by speaking with people. That goes for any company, but I think WooCommerce and WordPress specifically—so many people use it. They spend 12 hours a day using these products, building for them, using them as the end users, and they all have the same problems. They’ll all tell you the problems because they want you to solve them. It’s very approachable to go and figure out, “Okay, these hundred WooCommerce sites exist, and I’m just going to go email their contact email and say, ‘I’m building this product, and it might help you. Give me five minutes.’” A lot of people will, and you’ll also get nos, but that’s part of the process of building a company.

So speak with people who actually want to use your products, and you’ll find these common threads, and you’ll find where you could build valuable tools, and then you go and do it. Then you go right back to them and say, “Hey, remember that thing that you said cost you $5,000 a month? I just built something, and it will fix that for you, and there’s your value.” That’s been my approach with any product, and specifically with the AgentWP product. What are specific areas that we can solve with this type of interface, and how can we do it in a way that’s valuable for you? We find these really common pain points that can be solved in this specific way, and then we build for it. Ideally, we’ll be able to launch the core product that solves a lot of these problems and then continue to, one after another, cross off problems from this list and offer the solution for it. But the only way we know these problems is by talking with these people.

Tammie:
As we wrap, I want us to move a little bit to the future. We can move as far ahead into the future as you want to, but where do you see all these advances with artificial intelligence going in the future? What are the possibilities you see, or hope for even?

James:
Yeah, I hope that AI makes human creativity more valuable. I think we’ve already kind of seen that. I remember back in 2022 when the first AI craze was really popping up. It was something that everybody was horrified that it was going to replace art and replace human thought and all these different things. Obviously, I think there are risks there, but I also think that it’s serving to make the human aspect of creating a lot more important. So I think there will be disruption, and I think that we really have to… You can’t just stick your head into a hole and pretend and hope that it all goes away. I think things will change, but I think ideally, my vision for the future and the products that we’re building are products that enable the meaningful part of that work to be automated away and then really make the human aspect of it—specifically for me, like in agency and building WooCommerce stores—the things I loved were thinking about how we could design things and create these brands and this copy and these various elements of your store to make it convert for the users. So I hope that type of human interaction and the consulting, strategic, and thought parts really shine with AI and aren’t replaced. I think that is something that I’m already seeing a lot of people kind of going like, “Hey, AI is great, but it’s not a good artist, and it’s not a good copywriter. If you have 20 AI sites that all have the same copy, it all reads the same, it’s all the same mush.” I do hope and think that AI will surface the need for that human level of creativity and human level of interaction in web design and creation, but in everything, I think in every industry.

So that’s my hope for the future, and I do think that’s something that will be the case. I’m always a glass-half-full person, but I do think that’s something that will be the case. For us in WordPress, we’re trying to build those products that do make it easier for the builders of the sites to build the sites and to focus on the fun things, and to leverage these tools to better support their clients and to have their clients use these tools to get rid of those annoying 2 a.m. “make the logo bigger” requests. That’s our approach with everything, and I think in the future, you can’t just expect AI to disappear because I think it is a disruptor and I think it will change a lot of different things. But I also think that it’s a lot brighter future with AI tools and with these platforms to allow you as the creator to really focus on the creation and the architecture of a specific functionality or change, and the copy, branding, and human illustrations. I think that’s something that will be really important in the future. That’s what we’re trying to build towards. That’s what I’m hoping for in the future. That’s what I’m seeing early whispers of online and in digital media. “Hey, AI is good, but it spits out the same meaningless mush, so let’s not use it for that. Let’s use it to get rid of the annoying stuff and then focus on the fun stuff.”

Tammie:
You touched a little bit on fears, but you also mentioned your glass is half-full. So what are your biggest fears for open source and WordPress through the lens of AI?

James:
I think if you really want to take the horrifying view, you can see a future where a lot of creativity is just sapped away and replaced with meaningless mush sites, meaningless nonsense garbage content creation, and AI images that are all uncanny and look the same. I think that could happen if we allow it, but luckily I think the WordPress industry is kind of taking the charge on this with Matt’s focus on, “Hey, a WordPress site, an individual’s website, could become that individual’s expression of themselves online.” Especially in the future, it could become what it was like back in the day when a personal website was a way to express oneself. I think that’s something that it could grow back into. But I do think as an industry, we really have to be thoughtful and considerate of, “Hey, we don’t want mushy AI sites and mushy content and mushy images, where we can’t understand what’s human and what’s AI creations in the future.” I think the easy way to fix that is to just say, “Hey, let’s not do that. Let’s build products that enable us rather than replace us, and let’s go and really create meaningful human-like interactions online to show, ‘Here’s why you actually want this. Here’s an AI site, and here’s my site, and my site’s a million times better than this AI site.’”

That’s my approach, and I know I flipped it again into that glass-half-full approach, but that’s really something I think you have to do. You have to be excited about the future because it’s coming, and if you’re not excited about it and you just try to sit it out and pretend it’s all going to go away, it’s going to run you over, and it’s probably going to create the future that you don’t want. So for me, it’s like, “Hey, let’s build these products that do the things that we do want them to do and enable the things that we do want to be enabled, rather than let somebody else build a product that replaces me, or let somebody else build a website that is this mushy site when it could have been this human site, and now these clients don’t realize the value of what I could do.”

So I think you have to be mindful about what could happen, but I already see a lot in the Facebook groups, on Twitter, and in the social ecosystem of WordPress. I think people are really mindful of that, and if we try to build the future we want, it will likely be the future that we want. So I’m glad that’s already happening, and I think that will continue to happen into the future. Obviously, I’ll try to be one of the leaders in the WordPress space to go and build that future that we all want and build the tools that are valuable and helpful to everybody, not replacements or anything like that.

Jonathan:
James, I appreciate the care that you and the team have been putting into just thinking through things. There’s a lot that goes on behind the scenes. I know that’s something I’ve appreciated as I’ve been following along with what you’ve been doing. There’s a lot that goes into it to create those rails, to think about trust and safety, and I appreciate that. I’m looking forward to seeing how the next couple of months and years pan out. It’s great to have you in the ecosystem. You’re going to be at WordCamp US with us all in not too many weeks from now. If someone wants to reach out, how can people best get in contact with you?

James:
Yeah, if you want to reach out, I would actually love to meet people in person at WordCamp. I always see people online, and it’s really fun to meet people in person. So

find me on Twitter/X at @JamesWLaPage. You can send me a DM; the DMs are open. At WordCamp US and at any of the other events we go to—because we go to a lot of them now—if you see a tall, skinny dude wearing a WPAI hat, it is likely me, so come and approach me. I absolutely love being approached and just speaking with people and saying hello, so do that as well. But for any questions, ideas, partnerships, whatever, feel free to DM me on X. I’m always receptive there and would love to take the conversation a bit further. That’s what I do. I talk about AI, and I talk about WordPress. So love doing that. Don’t hesitate to DM me or say hello in person. I love seeing people.

Jonathan:
Excellent. Thank you, James. Talk to you soon.

Tammie:
Thank you.

James:
Thank you.