AI Can’t Learn Accessibility From A Broken Web

What happens when accessibility becomes a feature, not a fix? We sit down with Eugene Woo, CEO of Venngage, to explore how a design platform can bake inclusion into every step—from contrast-aware color pickers to exporting PDF/UA files that pass compliance without a remediation gauntlet. Eugene shares Venngage’s origin story, the early pressure from education and government users, and the decision to lead with built-in accessibility even when the market wasn’t asking loudly.

We dig into common misconceptions that keep teams on the sidelines: the belief that “accessible” means boring, or that compliance always adds time and cost. Eugene reframes accessibility as a creative constraint that improves legibility and clarity, especially when the tool handles structure and checks in real time. Then we tackle AI. Trained on a mostly inaccessible web, today’s models can draft fast but still hallucinate compliance. Eugene explains how Venngage pairs generative speed with deterministic rules for headings, layers, and exports, keeping a human in the loop where quality matters most.

The conversation widens to content strategy. Organic traffic that once flowed to blogs is shrinking as AI answer engines satisfy queries without a click. Eugene offers candid numbers and hard-earned perspective on what’s still working: unique data, useful tools, and product-led content that solves real problems. Looking ahead, he predicts pro tools will stay hands-on and AI-assisted, while non-designer platforms shift to prompt-first workflows—“apply my brand,” “swap this image,” “ensure contrast passes,” “export PDF/UA”—handled by an assistant that understands both design and accessibility.

Subscribe for more thoughtful conversations at the intersection of design, accessibility, and AI. If this resonated, share it with a teammate and leave a review—your support helps more creators build work that everyone can use.

Chapters

• 0:00 Meet Eugene Wu And Venngage
• 2:35 Why Accessibility Became Core
• 5:30 Innovating Built‑In A11y Features
• 9:45 Business Impact And Market Reality
• 12:55 From Engineer To Product Thinker
• 16:15 Misconceptions Holding Teams Back
• 21:05 AI’s Limits For Accessible Design
• 26:20 Jobs, Productivity, And Human Roles

Transcript

Debra Ruh: 0:00

Hi everyone, welcome to AXSChat. I was actually waiting for Neil to start the program, and I remembered he's not here today. He actually had to go and help his parents. And so it's just Antonio and me, and we're excited to talk to Eugene Woo. And Eugene is the CEO of Venngage. And I do know about this program and I've known about it for a while, but I did, Eugene, struggle just a little bit of the pronunciation, the pronunciation of your company. So maybe we could start off by learning more about you and a little bit of the origin story with your company, too.

Eugene Woo: 0:35

Sure. Thanks for thanks for uh having me on the podcast. I really appreciate it. So the name Venngage is actually a contraction of the word Venn diagram and engage. It's a much longer story, but we decided to just mesh those two words together. Very hard to pronounce and also very hard to spell because it has two ends. It's in it. But how did we get started? So VenGage is a design company that's primarily focused on uh infographics. So we are a design tool very similar to Canva, but with accessible with core accessible features built in, which we can talk about a little later. But essentially, it started with uh with infographics. I had this interest in creating infographics, and I wasn't a designer. I'm an engineer by by training, a software engineer by training. And so this is almost 15 years ago. I wrote a program that essentially took your LinkedIn profile, the information from the LinkedIn profile of your API, and then it would automatically visualize all your career into sort of what what we now call an infographic resume. Uh, you know, it would have like a pie chart, it would have like a uh timeline and bar graphs of your years of experience. It looked it looked pretty cool. So that was sort of the genesis of Ven Gage, and obviously we've you know done several rewrites, and here we are today. I mean, we are you know, I would say one of the few uh smaller companies in the design graphic design space, you know, uh you know, with the you know, obviously with the canvas and the Adobe's around as well.

Antonio Vieira Santos: 2:09

So uh Eugene, while Deborah is on mute, yeah, I I can ask you where accessibility, accessibility journey fits in the purpose when you're building a brand.

Eugene Woo: 2:23

Yeah, so so why accessibility? I think that that that is a very interesting question. We came into digital accessibility, sort of in a, I would say, as a as a product company in a very traditional way. Like our users were asking us about it. So we still have a lot of users in the higher educational space and in government, uh, so you know, from federal all the way to local governments. We began to get, I don't know, maybe like five, this is so pre-COVID, so maybe like six, seven years ago. We began to get a lot of requests for accessibility features. So they would ask us specifically can we export a you know, accessible PDF? You know, can I get a color contrast checker in all, you know, if you you know, I don't know how much you know everyone knows about accessibility. A lot of those questions were being asked, you know, of us. And we decided, you know, we made a decision about five years ago when we were rebuilding our platform to essentially make the entire platform accessible and to focus on accessibility as one of its core features. And we did this for two reasons. One, we thought it was the right thing to do. And the other one, also because when we looked at the market, most of the, most of our competitors, even now, but definitely back then, had like zero accessibility features. So if you looked at the Adobe's and the uh, and this will be Adobe Express, uh, or or what do they call it, Overview Spark back then, and Canva, and a lot of these, I would say, non-professional design tools, none of them had any accessibility features. And so we decided to be the first design tool to have built-in accessibility features. And we still are in some cases, and we did you know innovate a lot of the features that you now see in a lot of accessibility tools. Sorry, a lot of design tools. For example, having a built-in color checker in the in the color picker. So when you you know you pick a color, do you get this color wheel or something? We were the first tool that actually had a color contrast automatically in there, so that you didn't have to go to a separate tool, put in the codes, get the color contrast, and then come back to your design tool. Like everything is done automatically as you're picking the color. Now Figma has it. I think Canva has it as well. I'm not really sure. But we were the ones that innovated and you know, we've had that feature for like two and a half years, almost three years now.

Debra Ruh: 4:41

Wow. So y'all were leaders in the design of accessibility. I I remember really, really long time ago. I've been in the field such a long time. So this must have been like 2002, 2003. I had a booth in the United States at a big show. It was a Section 508 show, which 508 is accessibility in the United States, and Adobe was next to me. And poor Adobe. Everybody was lining up to yell at Adobe about not being accessible. I just watched this poor man just all day long. They just was brutal. But I know Adobe's been trying, but it's a really they have really big design tools. And so, um, first of all, thank you for taking such a lead on it. But I I would be curious because you're I was gonna ask you a question whenever I was on mute and and Tody will beat me. But I'm glad because I would say he took it in an interesting direction. I was gonna ask you this question, which I will now ask you, but um, you were a software engineer, and yet now you're actually a designer. So I'd be curious how you think you're looking at the world differently being trained as a software engineer and moving in design. But at the same time, I can't help but be so curious with this decision you made to listen to your customers and make sure that people were powerfully included. Has this helped your business in ways that it surprised you?

Eugene Woo: 6:06

So I would say it has. So I would say, yeah, it did open up a new market first. Although a quite a small market, I I don't think that the assessment market is uh, you know, at least for us, it's about it's it's still less than, it's still maybe 5% of our of our revenue. It's a very small part of our, although, although it is a big part of the platform, it is a small part of our of our business. The people who actually come to us and say we want accessibility and we want to use the accessibility tools, that is still that still remains a small part of our business, but it is growing. It is growing, and you know, we're we're quite happy with that. To answer your first question about transitioning from an engineer to a designer, I have to admit I'm still not a designer. I don't consider myself a designer. I don't, but I have had to transition from actually doing all the technical work to more of like a product designer, which is stepping back, thinking about what talking to customers, figuring out how looking at user experience as opposed to the code, you know, just how do I implement something? So I would say that has been pretty a pretty interesting experience for me personally, because now you're talking to customers, you're getting feedback from them, and not all of them are good feedback, there's a lot of bad feedback as well. And as you know, most engineers probably don't like to talk to customers, uh, me including, but but uh but I've learned how to do that, and and it's been quite rewarding because I learned a lot about you know, I learned something every time I talk to a customer.

Antonio Vieira Santos: 8:05

If you look at your customer base and the people that talk on the daily basis, what are some of the preconceptions that marketing teams still have in what concerns to scale content that is accessible on their campaigns, on their products, on everything that they do?

Eugene Woo: 8:25

I would say the the number one misconception is that accessibility is still very, very difficult. Accessibility is not accessible, I would say, because when you look, at least for graphic design and and PDFs, if you look at the status quo right now, is you create it, you take it to uh uh Adobe prof uh Adobe Pro Acrobat, and then you have to remediate it, which is you have to like fix it. And it's a very technical process. You you know, even even I'm not even I don't know how to do it that well. I can like kind of like do it a little bit, but not really. And I and I've been in this PDF accessibility field for for years now, and I still don't know how to do it. So I think number one is that there's this mental barrier to go, like, this is really difficult, right? Like, I am not a professional PDF remediator, I cannot do it. And so it doesn't scale. And then I think the other misconceptions is that it's really hard to make something accessible and also make it aesthetically like beautiful. So there is that thing about oh, the accessibility police is gonna come in and essentially make my beautiful design really boring. Why don't I just type out a Word document? You know, that will be accessible and forget about all this other design stuff. I'll just give you a pure text, then it'll be accessible. So I think there's also that it that's an exaggeration, but but there's also that misconception that, oh, I cannot apply my brand colors properly. I cannot have like a aesthetic flair and design because it's not accessible. The accessibility team is gonna come down on me if I do that. So I think that's also a misconception, uh, largely because they don't, you know, because of the first misconception, they don't understand what it takes, you know, to make something accessible or what the what the um um what what are all the criteria for for accessibility? So I between those two, like there's this big mental block, and like how do you know how can I even go into accessibility?

Debra Ruh: 10:20

Okay, I I gotta ask the the the word that then I know you've nailed all this. How is AI changing um that what you're doing only because it's changing the way we create content? I don't know if y'all know this now, but you don't know these days uh if a video is real or not, if the content's real, right? So it's like everything has changed. And at first it seemed like we went to AI, big corporations throughout their marketing people, and then they're like, oh no, no, come on back, because actually we do do need humans, but are there risks associated using AI-generated visuals? Yes. I it I mean, it's such a big topic, Eugene. How do you begin to wrap your hands around? And of course, I know y'all are seen as leaders in the AI in this field. So and of course, oh, we were saying this before we went on air. I think we all know here AI is the continuation of the internet and technology. But anyway, so Eugene, how's it going to change everything? Because it's really interesting.

Eugene Woo: 11:18

Well, yeah, I I think I'm gonna talk through it as it pertains to design and accessibility. As it pertains to design and accessibility, AI is still fairly uh nascent. It's still in infancy right now. And and I think um, you know, I I was talking to someone else like two weeks ago. I think the the main issue is AI, the the image generation uh models and the language, the LLMs, the language generation models are trained on existing data, right? Existing websites. And most websites, as you know, are not accessible. So when you ask the AI to create a website for you, well, guess what? Even if you tell it please make it accessible, the examples it's been trained on are still not accessible. So it will try its best. But because of its data available, and the I think the web in the organization that does this, uh, that does the survey on the top, I think, one million websites, it's still like 90 something percent of the websites are not accessible. It's an astoundingly large number. So I think that because of that problem, AI is helpful, very helpful in building up the scaffolding and all the stuff that takes a lot of time. But as it pertains, and can make it quite beautiful, but as it pertains to accessibility, as of now, it still doesn't have enough data, unfortunately. And we've tried it, and we would the LLM will sound very confident, the AI will sound very confident, will say, like, I have coded this website, it's completely accessible, the color contrast is great, and then I'll look at it and go like, no, it's not. That isn't that is a lie.

Debra Ruh: 12:50

So true.

Eugene Woo: 12:51

Yeah, Carl Groves actually he's uh he's uh the accessibility expert. He he made this quote and he said something like, Okay, you know, Carl, he said that AI is as bad as it is right now, which is the truth, right? So it's only gonna get better. So I do think that as bad as bad as it is right now in accessibility, it's gonna improve. And so I'm quite optimistic that AI will be able to help us create accessible websites. We we just got to put more data and train it properly. And as it pertains to graphic design, it should be able to also create things that are accessible with color contrast, proper color contrast. It doesn't do layering right now, but that's what we do in Van Gage. We actually take the A, you know, we mix AI with our own, with our own technology, and we'll be able to create different layers with proper headings and and and all of that. And so I think that, you know, I think that AI will be very positive, you know, with with as it pertains to digital accessibility, although it is still at its infancy right now.

Debra Ruh: 13:52

I agree. And before Antonio comes in, I'm just gonna say too that I I know that there's been talk about AI is gonna wipe off the accessibility field in a year or something like that. I think instead the opportunity is to really drive in as a you know entrepreneurs and think about how AIs can improve what we're doing. They are tools. AI is a tool, and and it can solve so many problems, not all of the problems, but I think what I'm seeing, and I I think you would you would say this as well, Eugene, it's like AI used as a tool can be so beneficial to our community. But there, I think the accessibility providers, the vendors, they need to stop fighting AI and figuring out hey, AI could really help. Instead of just assuming AI is gonna kill us all, I think we have an opportunity to look and see how do we do it different? And I think you are doing that, Eugene. So compliments to you there.

Eugene Woo: 14:44

Oh, thank you. Yeah, and and I yeah, and I agree with you on that statement. I do think that, I mean, it's not just AI, they're saying it for every other industry. You know, they're like, well, all white colored jobs are gonna be gone, for example. Right. Uh and and I'm like, I don't know. I don't think so, you know.

Debra Ruh: 15:02

We actually have to, the humans have to have the will to make that happen. I hear leaders saying, uh, I heard this about six months ago, so I guess we've got a year and a half, 90% of jobs are gonna be gone. And I was like, are you kidding me? We can't even figure out how to do our yeah, draw the littlest things right now. Right. Humans would have to agree we're gonna wipe all these jobs out that quickly. It doesn't even make sense. Instead, looking at it as a tool, looking at it, how can we, with everything that the AI is doing with our graphics, how do we make sure it does work? And we still need human engineers and design in this.

Eugene Woo: 15:38

Yeah, and I also think the the other thing, well, I'm well, I don't know, maybe AI will figure out how to do this, but you still need humans to put new information into the models, right? Like AI is AI as AI as we know now, it's not we haven't reached like AGI or like the actual intelligence. So we still have to feed the models with data that are biological created by humans. I think until AI creates its own data and its own original research and all of that, which may happen. I don't know, maybe that will happen. I think that will be a long, you know, that may be decades. Who knows? I don't know.

Debra Ruh: 16:12

And the I still think and the bias that we've scheduled in.

Eugene Woo: 16:16

Correct.

Debra Ruh: 16:17

Yes.

Eugene Woo: 16:17

Correct. Yes, and all of that stuff. And as it pertains to jobs, I would say that, you know, so I'm you know I have a company, a small company, we employ about like 50 people. And I always use myself as the when I hear about these statements, like oh, all the jobs that can be gone. I always go, like, Am I able to now replace a person with AI? And the answer as of today is still no. I have not been able to replace, yeah. I haven't been able to replace, they've made people a lot more efficient and more productive, including myself. It's made me a lot more efficient and productive, so I'm able to do more. However, it hasn't actually replaced, I haven't been able to replace anyone. And I always joke, I'm like, I wish it were true. Then I maybe I don't I can have fewer employees, right? But truth is, you know, it's not true. Yes. And I I think and and the people who said the the who made the predictions of like how I think they made those predictions a year ago that said, oh, in a year they'll be and it's like it that prediction keeps going on until um I don't know, until it's not true, I guess.

Antonio Vieira Santos: 17:14

Yes, yes. When when I was a kid, I remember having a Walt Disney book, and on that book there was a factory building robots. And no, and no, this was many years ago, and the the robots were being built to replace all the cartoons, you know, and then you start to have all these issues where the robots are going to fight against themselves, fighting against each other. So these ideas of robots replacing humans, this is something that has been around for quite a number of years around science fiction, and now things are becoming more you know into our day-to-day reality, and but these old images are still coming over and over, repeating themselves. I have a question, Eugene you are an entrepreneur, you have you have responsibility over people. You probably invest money on personal development, your your employees, you you might invest on research. With all this pressure on AI, don't you feel that sometimes oh we all need to rush on AI, we need to do this, this, the AI. We might leave behind other things that are also important.

Eugene Woo: 18:25

Oh, that's a very good question. So, yes, I agree, I agree, I agree with what you're saying, and we are probably um guilty of doing that. We are so focused on AI right now. Uh, our I would say in the last 18 months, the the main user experience of of designing uh uh infographic or a poster used to be you pick a template and then you kind of like manually do it, manually like swap things around. Now it's all you write a prompt. You pick you can still pick a template, but all will pick it for you. But you write a prompt, you tell us what you want to want to build or what what what you want to like design, and then we just like do it for you, and then there's a chat assistant. So you're right. We have been very focused on AI, but but I do think the the main thing is it's important not to do AI for the sake of doing AI. So we're always, you know, and going back to user, right? Like so for us, our our our problem that we're trying to solve for our customers and for our users is how do you make design more accessible in in all in not just digital accessibility, but also like make it easier for non-designers, people who are not professional designers. And that has always been the focus. So so even when we're thinking about AI tools, we don't just like, oh, let's just implement an AI feature just because it's out there and you can do it. We always figure out, like, hey, what's the end goal? Like the user, you know, the user still needs to do XYZ. And if AI can't do it, well, we'll you know, we'll we'll let them do it manually still. For example, picking a template, like a lot of a lot of people still like to browse and pick, as opposed to have the AI uh, you know, choose it for you. And so we give them both options now. Like you can, you know, we can pick one, and then we'll have a bunch of, we'll have the gallery there, and you can pick another one, and then we will, you know, we will we will do that. So I think it's and so you're right, it's very important to still remember like let's solve human problems and use AI to help us, as opposed to just do AI for the sake of AI, right? It's this is sort of similar to how when like when uh what is that, when crypto was out there, people just building crypto, just building crypto for the sake of crypto, right? Oh, I'm gonna build this Ethereum back platform for chat. I'm like, well, you can just build a chat platform if you wanted to build it. There's no need to add like a you know crypto, a Bitcoin backend to it or something like that, right? So I think the same with AI, just don't have it for the sake of having AI. It is all about helping us solve problems at the end of the day.

Antonio Vieira Santos: 20:52

So to complement this a little bit, we we know that today many people start their search within the you know, in ChatGPT, Perplexity, and many other platforms. Uh, have you felt the need to do any changes in the in the way how you communicate, in the way how you build your site, to adjust to the way how people are changing behaviors in relation to search?

Eugene Woo: 21:17

Yes. I would say so. Search is the main method that people find us, or what we call organic search. And our blog used to be the number one place where people find us. It's not, but as of you know, I would say last year and and uh and and continuing now, the blog is like 10% of the traffic that it used to have. So we used to drive literally millions of visitors to our blog. Now we drive, I don't know, like less than a hundred, like 50,000, maybe. It's pretty much been like wiped out, pretty much by the uh by the uh by the AI uh answers. And so we have been trying to figure out how do you how do we make our content more you know more mentionable? But I think at the end of the day, it didn't really, I don't know. I don't know this. I mean, if someone knows, they can tell me. We haven't actually changed that. We're still trying to figure out is there a way to quote unquote write content that can actually get someone to come to our blog? And I think we it's it's a really difficult battle because even the content we write, the the AI, you know, the LLMs will scrape it, we'll put it in their model, and they will display it. You know, it might give us a link, it might say, oh, this was mentioned, but people don't click into it, they just read the information. There is no need, right, to come to our website when they can read it. And I do it myself. Like I'm asking ChatGPT everything, and I'm just reading it. Very rarely do I ever click on the on the references, only if I have to see, well, it used to be only if I have to see an image. Now they are showing images as well, so I don't even have to do that uh anymore. And so so I think it it is a big problem in a sense that content creators really it's very difficult to get a visitor uh to come to your site now because all that information is available on the on the on the on the AIs with Perplexity and Chat GPT and Gemini. There's no need to. And yeah, I mean, I don't know the answer to that question. I wish I did. Then we would get all that traffic back.

Debra Ruh: 23:13

That is so interesting, though. I haven't heard anyone talking about the numbers like you have. I'm also a content creator. I mean, Access Chat is a content creator. It is really interesting though, Eugene. It is interesting because how do you stand out in this world? It's it's fascinating. And it also feels like sometimes the things that stand out the most, of course, are the things that are um a little bit controversial and going to get people a little heated. And which of course we have a uh, you know, that's just part of our world right now. But I'm gonna ask you a very, very unfair question, Eugene. I'm sorry, but uh, where do you think the future of design tools are heading in the next five to 10 years? And it's so funny to say five to ten years because it's like in the next five months. I mean, it's we used to say five to ten years. I don't even know how that works anymore, but yeah.

Eugene Woo: 24:07

So, so I think the so I'm gonna talk about I think there's two tiers of design tools. So there's one would be like we'd say it to professional design tools, these will be the Adobe Illustrators and all of that. I think those design tools will pretty much be the same, but but AI powered, right? So you are now instead of you having to draw the actual paths, you can just like either you know dictate or just type it in and say, you know, draw all these paths and then apply these filters, this mass, and all you know, all the features that you could do. I do think those, I think the incumbents will pretty much still be there. That that is my on the professional side because because the barrier is so much higher, it's a lot more difficult to do those kind of stuff. On the lower end, on on the non-professional side, so these are tools like Van Gage and Canva and the Abobi Express, uh, the easy to use or the do-it-yourself kind of uh uh tools. I think those tools would will change, would be very, very different. I think there will a lot there will be probably all be prompt-based tools where you're just prompting either your speaking or you're typing, and you're not manually editing as much anymore. So the what we call the drag and drop or the WYSIWYK, you know, what you see is what you get interface, uh, which is the you know the PowerPoint interface where you're dragging things around. I think that interface will probably change to something that's more AI friendly, which at least for now is the prompt. It's the it's the chat, right? For now, it's a chat. So you'll have this AI assistant that is gonna just do the work for you, as in like, oh, apply my brand colors, change the change this image to something else, apply a filter to this image. I want it to say this instead of that, and it'll just all do it for you. And you can kind of speed up design a lot faster that way. And then when you're done, you're you can say something like, oh, make sure it's assessed, you know, make it everything accessible, make sure it's accessible, export it to an accessible PDF format, and so forth and so forth. Like, which that is actually quite close. I think we're probably like a year away from that, six months to a year away from that. Um, I mean, even in Vengage, we are building our AI assistant, our chat assistant, and it can do you know a fair bit of things right now.

Debra Ruh: 26:20

Yeah, it's a it's amazing how fast it's changing and moving. It's hard to predict it. Um, but I know that we've kept you on for 30 minutes. So first of all, I want to thank Amazon for sponsoring us. We are so grateful for them keeping us on air. But Eugene, I want to make sure that people, first of all, we'd like to know how to read that blog because we should read blogs written by humans, especially when it comes to design and accessibility. So please do go and read this blog because uh those numbers are scary. We need to read blogs by rural humans. So, but will you tell the audience how to find it about your blog and also your company and how to contact with you? Sure. Contact with you too.

Eugene Woo: 27:01

So the company's vendage is V-E-N-N-G-A-G-E.com. And once you're on the home home page, you can click on our blog. And again, we are we are uh we are a design platform very similar to Canva, except if you want accessible uh designs, especially in accessible PDFs, we have all those features built in. You don't have to remediate your design after you're done. Pretty much when you export it, it is compliant. And uh yeah, and we're probably the only tool that does a PDF UA, which is the standard for PDF compliance out there right now.

Debra Ruh: 27:33

Right. And and I also would say to the audience, we need to support vendors that are supporting us. So I actually use Canva and some of the Adobe tools. I need to come over and use Vingage because seriously, one thing we have to do as a community is we have to support the vendors that are supporting us. And so that's one reason why we wanted you to come on, Eugene. So we thank you for that.

Eugene Woo: 27:55

Yeah, and and on that on that point, I was gonna say like we do offer uh free, sort of like lifetime free accounts to to consultants like yourself, consultants and and anyone in the accessibility space. So just contact me. I'm on LinkedIn, by the way. You can contact me, Eugene, with a LinkedIn, and uh we'll give you a free account.

Debra Ruh: 28:14

Yay, yay, all right. Well, I will be doing that, Eugene. Um, thank you so much, Eugene. Once again, the website is www.v-e-n-n-a-g-h.

Eugene Woo: 28:25

G-A-G.

Debra Ruh: 28:26

G-A-G-E. It's just like it's it it so thank you, Eugene. Thank you, Antonio. And Neil, hope you're having fun with your parents. Bye, everyone.

Eugene Woo: 28:34

Thank you, thank you, Deborah. Thank you, Antonio. Appreciate it.