With the growing popularity of vibe coding, I decided it was time to pressure test it with my own accessibility workflow. Inspired by Mel’s example, I decided to create a simple one-pager for a nonprofit leadership conference. In this blog post, I’ll share what I learned.

TL;DR Automated accessibility checkers don't replace real testing with assistive technology (AT) such as screen readers and magnifiers, preferably by someone who's a natural user of AT. Also, AI will confidently tell you that something passes an accessibility check when it doesn't.

Monday: First Prompt

1:29PM – I give Claude our existing source information, visual style guidelines, and ask it to create a one pager with WCAG (Web Content Accessibility Guidelines) 2.2 AA conformance.

I spent about three painful hours of prompting to adjust the copy and formatting over and over again. Even though I know exactly what I wanted at a code level for the HTML and CSS, the process of making these adjustments is much slower in a prompt-based interface. It feels like standing over someone’s shoulder and telling them what to do rather than just doing it myself.

4:41 PM – I realize that Claude has gone off the rails when making decisions about other adjustments, so I have to force accessibility back on track again. My goal is for this accessible PDF to be available at the conference in print with a digital version on my website, so I carefully check font size and type, readability, color, and contrast.

6:13 PM – Something unexpected happens. Claude decides to inject code to change color and contrast, but only when the PDF is printed. The digital version remains the same.

6:34 PM – After two hours of unsuccessfully trying to get the HTML converted to a printed PDF with the correct margins, color, and contrast, I stop and call it a day.

Tuesday: Try Again

10:57 AM – I beg Claude to help me make this PDF print correctly. No matter what I tell it to do or how I adjust the print settings, it loses a footer, prints with a gap on the edge of the page, or a number of other quirky issues.

11:08 AM – I give up again and revert it back.

Wednesday: Last Chance

1:00 PM – I have some feedback on wording and format and I’m bracing myself. At this point, I’m tempted to pull what I have into another authoring environment, but I decide to stick with it. I give Claude some prompting to continue tweaking the printed version.

2:51 PM – After nearly three hours of adjusting padding in different areas of the document, I’m finally getting a viable printed version. Also, at some point along the way, the issues I’d been having on print mysteriously disappear. I have no idea how it fixed itself and it couldn’t tell me.

Results

First, I print the full color one-pager for the conference and it looks decent enough. Are there still things I’d change? Yes.

Am I done dealing with vibe coding? Also, yes.

Next, I load the digital version of the PDF to our website. Remember that Claude created two versions of the PDF: one that was supposed to be optimized for print and one for digital use.

Time spent: Approximately 8 hours, excluding print time (design layout and accessibility only; we already had the style and content)

Water used: Based on estimates for intensive AI vibe coding sessions, 8 hours of prompting Claude to generate a one-pager could consume approximately 4 to 18 gallons of water. Reference: Hidden Costs of AI: Practical Steps to Reduce Energy and Water Use

Review – Digital Accessibility Checker

Claude promised that the digital version of the one pager passed WCAG 2.2 AA with no issues. WAVE gave the digital version of the one pager a 10 out of 10.

Review – Human

The source HTML version that Claude created was okay. It was missing alt text for the logo and some of the headings were not tagged properly. Reading order was correct. Unfortunately, the HTML version wasn’t what I needed. I needed an accessible PDF to use in both print and digital formats.

The printed PDF met acceptable accessibility standards. It was high contrast and when printed and the text size was acceptable.

Digital accessibility of the PDF is mediocre at best. The Build Capable logo doesn’t have alt text. There are missing heading tags and some of the heading tags break one heading into multiple headings (each word is tagged as it’s own heading). The reading order is not retained in the accessible PDF version. These would all be easy for me to fix if I could edit the code myself, but Claude just wasn’t getting it done.

Here’s a VoiceOver (Mac) screen reader demo comparison, starting with the source HTML version and ending with the digital PDF.

This activity demonstrates that Claude falls short when it comes to putting a human in control of the design. I found vibe coding a simple design very frustrating and ended with an output that I needed to edit in a different tool where I could really control the look and accessibility functionality.

The opportunity for AI in the design space, I believe, will be focused, AI-supported authoring environments that can quickly provide initial structure, but give the human direct control of editing the elements. I’m certainly not convinced that a chat-based authoring environment alone is an efficient, effective way to do this work – especially not for experienced design professionals. It may get you started with a fast first draft, but cross your fingers that it’s one you can work with.

As the world’s leading aviation services provider, Swissport supports 65,000 colleagues across 45 countries in some of the most demanding operational environments in the industry. Ramp agents, cargo handlers, and passenger service staff don’t sit at desks. They need training that meets them where they are, and insight that goes beyond a ticked box.

So when it came to rolling out Year-End Review training globally, Swissport took an innovative approach. They captured rich behavioral data in real time using xAPI-powered, link-based delivery through Build Capable XCL that reached every colleague regardless of location or LMS access.

What they discovered reshaped how they think about learning design.

Learners navigated content selectively, returning to videos as just-in-time reference tools, and engaging in ways that reflected operational behavior rather than passive course consumption. Managers showed distinct engagement patterns from general employees. The data surfaced process improvement opportunities that went beyond the learning program itself.

100% of the target audience was reached. Decisions were grounded in observable behavior rather than assumptions. And Swissport now has a repeatable model for delivering and continuously improving learning at global scale.

This is what it looks like when a world-class organization treats learning analytics as a strategic capability.

Access the full case study, How Swissport Advanced Workforce Learning Through Next-Generation Analytics, from Daniel Lever, Global Head of Academy & Learning Technology and Paul Griffin, Global LMS Manager at Swissport.

• The administrative processes were manual and time-consuming. 
• Engagement was low, with fewer than 25 learners per month.
• The pricing model made growth difficult. More users meant higher costs.

Midmark RTLS took a different approach with Build Capable XCL. Instead of forcing learners into a traditional LMS, XCL enabled content to live where work already happens. Training was embedded directly into their knowledge base in Zendesk, eliminating the need for manual enrollment and improving access.

The experience became intuitive and seamless. Learners could start and stop without losing progress, engage with interactive content, and access what they needed when they needed it.

From an operational perspective, the shift reduced administrative burden and provided real-time insights into learner activity and performance. The flat pricing model also removed cost barriers, making it possible to scale to thousands of users without increasing expenses.

This case highlights a strategic shift from platform-centric training to an ecosystem-based approach.

Access the full case study, Midmark RTLS Modernizes Channel Learning with Build Capable XCL, from Susan Weyandt, Partner Development Manager at Midmark RTLS.

Build Capable is the creator of XCL, a platform that enables organizations to deliver and track online learning experiences at scale without the constraints of an LMS.

Yet Analytics is the creator of SQL LRS, the most widely installed open-source Learning Record Store (LRS), distributed under the permissive Apache 2.0 license.

Together, these technologies form a modern, flexible learning data stack built for interoperability, analytics, and AI-driven applications from the ground up.

“Too often, the LMS becomes an expensive way to solve yesterday’s problems,” said Yet Analytics CEO, Shelly Blake-Plock. “In the AI era, the focus shifts to performance, interoperability, measurement, and AI-ready systems that actually evolve.”

Beyond their respective platforms, Build Capable and Yet Analytics bring deep expertise to support organizations in:

• Designing custom learning experiences and implementing xAPI
• Building robust Learning Record Store infrastructure and data architectures
• Pre-processing xAPI data with conditional logic for seamless integration into AI and analytics workflows
• Developing end-to-end strategies for organizations building or scaling modern learning ecosystems

“Organizations are sitting on valuable learning data but struggle to capture and use it,” said Build Capable CEO, Sarah C. Mercier. “We’re excited to partner with Yet Analytics to change that and make learning data part of meaningful business intelligence.”

This partnership represents a shift toward a more open, data-centric approach to learning. One where experiences are not confined to a single system, and where learning data is ready for advanced analysis and AI from day one.

Ready to move beyond the LMS and unlock the full value of your learning data? Connect with Build Capable and Yet Analytics to design a modern, AI-ready learning ecosystem built for what’s next.

Build Capable XCL: The End of Platform-Dependent Learning

We recognize that flashy platforms and tools get a lot of attention, but these rarely reflect business needs in the real world. We’ve created a groundbreaking learning technology that is a simple, elegant solution to address gaps in existing learning ecosystems.

Our goal was to improve access to a wide range of learning experiences while collecting rich data to support business needs at scale.

Unlike rigid platforms, XCL enables “headless” learning: flexible, branded, and easily deployed where learners already are. Whether you’re a nonprofit, enterprise, or educator, XCL empowers you to deliver learning where it counts and measure what matters.

“This Gold winner demonstrated how it’s possible to innovate and solve a long-standing industry problem. XCL’s impressive, accessible, and disruptive tool plugs into existing tech stacks to connect any digital learning experience to analytics. It challenges what a platform is, how it works, and how data in learning can be leveraged powerfully.”

– Learning Technologies, Awards Announcement

Susi Miller, who was there to accept an award for Excellence in the Design of Learning Content, graciously accepted our award at the gala in London. Meanwhile, the Build Capable team celebrated from the DevLearn Conference and Expo in Las Vegas. As fellow accessibility advocates, we are grateful to Susi for representing our team. These awards demonstrate that accessible design isn’t a barrier to innovation — it’s the foundation for it.

What is Build Capable XCL?

Build Capable XCL is a patent-pending tool that transforms digital content into a trackable link and QR code. No LMS, logins, or per-user fees required. From legacy SCORM packages to virtual reality simulations, XCL translates a learner’s actions to xAPI data, collected in real time. It’s fully accessible, works at scale, and dramatically reduces costs.

By removing the friction of traditional platforms and enabling flexible, accessible, data-rich experiences at scale, XCL changes how organizations think about their learning ecosystem. It isn’t just new, it’s necessary. XCL offers real-world capability, empowering everyone to share learning experiences and collect data that matters.

What is Build Capable XCL?

XCL is a patent-pending tool that redefines how learning experiences are delivered, tracked, and analyzed. XCL enables organizations to create persistent, shareable links and QR codes to digital content, from SCORM/xAPI packages to PDFs, videos, games, websites, and more. These Learning Links can be distributed anywhere: email, SMS, Teams, Sharepoint, Slack, LMSs, intranets, QR codes on physical products, or even printed materials.

When someone interacts with the Learning Link content, XCL instantly translates that activity into Experience API (xAPI) data and reports it to a learning record store (LRS).

What sets XCL apart is its simplicity, accessibility, and flexibility:

• No user accounts to manage
• No rigid platform requirements
• No per-user fees
• Accessible to assistive technologies such as screen readers

From global nonprofits distributing job aids to enterprises analyzing real-time training simulations, XCL offers a cost-effective, future-ready alternative to traditional learning platforms.

XCL: Designed for Accessibility

Accessibility has always been at the heart of Build Capable’s mission. As a leader in learner-centered design, we believe accessibility is a requirement, not an afterthought. This achievement reinforces our commitment to creating inclusive learning technologies that remove barriers for everyone, from XCL administrators creating Learning Links to individuals accessing content through them.

“Historically, the tools and platforms we use in learning and development have not been accessible. It’s time for that to change.” said Sarah Mercier, CEO of Build Capable. “XCL passing WCAG 2.1 AA standards validates our team’s commitment to accessibility and ensures our solutions are usable by everyone, including those who use assistive technologies.”

Read our accessibility statement and download the full Accessibility Conformance Report (VPAT) on the Build Capable XCL Accessibility page.

From an accessibility perspective, however, confetti explosions are not great for everyone – the motion can adversely affect some individuals.

The good news is that you can have your confetti and accessibility too!

The way to implement confetti fun and stay committed to creating accessible digital experiences is to provide folks with a way to turn it off. In Build Capable XCL, we added a simple toggle button in the account settings.

This applies to online learning experiences, too. Recently, Melissa Milloway shared a portfolio project where she implemented confetti. The effect is super cool and strategically used once at completion, but she also provides the option to turn it off.

So, go ahead… use confetti. But make it accessible.

But wait. What about situations like these?

• Training for an external audience, such as nonprofits sharing educational materials with the public (where “seats” in an LMS aren’t practical or affordable)?
• Organizations that need to train frontline workers and contractors (where you need a quick attestation without paying for user accounts)?
• Research organizations that need to collect training data over a short period of time (where implementing an entire platform just doesn’t make sense)?
• Providing performance support such as job aids, videos, and chatbots?
• Delivering and tracking anything other than SCORM-packaged eLearning?

The learning management system (LMS) model has always had its limitations, but we were forced to either cram your company into one or set up your training content manually across disparate systems. This results in a poor learning experience for people, but also means your data is all over the place. Until Now. Enter Build Capable XCL.

XCL started as a concept for our clients who weren’t good candidates for a traditional LMS. They wanted flexibility in delivering their content and they wanted data. They also weren’t ever going to shell out the money for a learning experience platform (LXP) due to the nature of their work or number of learners. Nothing existed to meet their needs, so we created XCL to help!

What is XCL?

Build Capable XCL is a tool that creates a shareable link to your content. When someone uses the link, you collect data about their actions in real time. You can share these XCL Learning Links with as many people as you want since there are no user accounts or per-user fees. Learners don’t have to create an account or log in to a platform – they get direct access to your training and performance support content.

Part of the magic behind XCL is the xAPI Translator. XCL translates actions taken by a learner to an xAPI format and sends that data to a learning record store (LRS) without you ever writing a line of code. This is your “easy button” to get super detailed data from content like videos, PDFs, eLearning, and more.

XCL’s DevLearn DemoFest Debut

We teamed up with the folks at RISC to create a solution for one of our clients where we paired XCL with their tool, xACTION. In the project, we used XCL to deploy and track a variety of training content for a nonprofit with a public audience. Meanwhile, xACTION sat in the background “listening” for data to come through. When certain data rules were met (e.g., completion of several eLearning modules and videos), xACTION emailed the learner with their certificate and recorded their completion on their transcript.

Pairing these two tools allowed us to replace the functionality they would have gotten with an enterprise LXP at a tiny fraction of the cost – and the learner never had to log into a platform. This solution is a huge win for our client, and the attendees at DemoFest agreed!

We took home the award for Best Performance Support.

But, just as important, we got lots of great feedback and ideas from folks. It was incredibly helpful to learn where attendees thought we absolutely nailed this solution. We also left with some new feature ideas to include as we wrap up our beta release!

What’s Next?

The nerdy part of me is super excited to see all of the innovative ways that the learning community will use XCL. It saved our team days of custom integration work that I can now do in seconds. Our Build Capable clients save time and money so they can focus on their organizational goals. Everyone wins.

If you’re curious about whether XCL can help you, you can try it free! Head over to Build Capable XCL to learn more and sign up for Beta.

Do you wish that you could write xAPI statements faster? Maybe you’re not a JSON pro and struggle to write or troubleshoot custom code to use in tools like Articulate Storyline? Or, maybe you want to construct a statement triggered using JavaScript?

Now that custom GPTs are available to use with the free ChatGPT account, I’m sharing our custom GPT, xAPI Statement Constructor to help.

Image description: The xAPI Statement Constructor GPT chat landing page. There are four conversation starters: What is xAPI? What is cmi5? How do I add an xAPI statement to Storyline? What xAPI statements does dominKnow send? You can begin a chat by selecting a conversation prompt or using the chat entry. There is a disclaimer at the bottom of the page that says, “Build Capable workspace chats aren’t used to train our models. ChatGPT can make mistakes.”

Behind the Scenes

The xAPI Statement Constructor is trained on the ADL xAPI wrapper. It works best when providing a prompt that includes a list of parameters. Here’s an example starter structure (credit to Wendy M. Morgan), which I’ve tested numerous times and it consistently gives me a solid result:

• Verb: What action is being performed? (completed, attempted, passed, failed, etc…)
• Activity Name: The name of the activity or task. For example, this could be a course ID.
• Activity Description: A brief description of the activity or task.
• Result: The outcome of the activity, if applicable. (score, duration, etc…)
• Result Success: Whether the activity was successfully completed (true/false)
• Context – Parent: The parent activity or broader context in which this activity takes place.
• Context – Grouping: Any grouping information that links this activity to others.

Note: Some parameters, such as context, may not be needed.

Generate a Statement!

To demonstrate how this GPT works, I’ll create an xAPI statement that is triggered when satisfying a competency by completing a course. For example, I am currently facilitating the Learning Technologies Certificate Program for the Association for Talent Development (ATD). When an individual completes this program, they increase a capability that is part of The Talent Development Capability Model in the Developing Professional Capability Domain: Technology Application.

Let’s say that ATD uses xAPI to track an individual’s progress along the TD Capability Model. In this case, they could trigger an xAPI statement to be sent to an LRS when an individual successfully completes this certificate program, thereby increasing the ‘Technology Application’ capability. Here is a prompt with some possible parameters for the statement:

• Verb: completed
• Activity Name: LTCERT0603
• Activity Description: Learning Technologies Certificate Online
• Result: capability increased
• Result Success: T
• Context – Parent: Technology Application
• Context – Grouping: Developing Professional Capability

You may notice that I did not include the actor in these parameters because my assumption is that you will pull that information elsewhere to complete the statement rather than hard coding it (e.g., Storyline can grab the actor data from launch). However, you can prompt the GPT to include a generic (or specific) actor if you choose. Just include the actor parameter and details in your prompt list.

Image description: This is the xAPI Statement Constructor GPT chat. I have pasted and submitted the prompt with details directly in the chat with no other instruction and it responded with the xAPI statement JSON. This is the default format that the xAPI Statement Constructor provides, but you can change this by modifying the prompt format I shared.

Image description: The xAPI Statement Constructor wrote JSON based on the parameters provided. There is a feature to “copy code” in the top right corner of the code dialogue box. The GPT chat also states that the statement reflects the completion of the ‘LCERT0603’ activity with an increased capability result and success marked as true. The context includes both the parent activity ‘Technology Application’ and the grouping activity ‘Developing Professional Capability’. JSON text:

{
  "verb": {
    "id": "http://adlnet.gov/expapi/verbs/completed",
    "display": {
      "en-US": "completed"
    }
  },
  "object": {
    "id": "http://example.com/activities/LTCERT0603",
    "definition": {
      "name": {
        "en-US": "LTCERT0603"
      },
      "description": {
        "en-US": "Learning Technologies Certificate Online"
      }
    }
  },
  "result": {
    "completion": true,
    "success": true,
    "response": "capability increased"
  },
  "context": {
    "contextActivities": {
      "parent": [
        {
          "id": "http://example.com/activities/TechnologyApplication",
          "definition": {
            "name": {
              "en-US": "Technology Application"
            }
          }
        }
      ],
      "grouping": [
        {
          "id": "http://example.com/activities/DevelopingProfessionalCapability",
          "definition": {
            "name": {
              "en-US": "Developing Professional Capability"
            }
          }
        }
      ]
    }
  }
}

You can further prompt the GPT, such as replacing http://example.com/ with your desired domain and adding parameters.

Keep in mind that this GPT is going to be most helpful when you already understand how xAPI works and have established a data strategy. Tools like ChatGPT are effective when you already know how to do something and want to do it faster/more efficiently – otherwise you run a high risk of errors.

I’m excited for you to try the xAPI Statement Constructor! Let us know how it goes. We will use your feedback to continue to train and improve the GPT.

Need some help getting started with xAPI?

xAPI Basics

Being able to learn and participate in work and in society is a human right.¹ When we don’t create learning experiences that are inclusive and accessible to all, we reinforce barriers that prevent people from participating and therefore deny them an essential human right. It’s a vicious cycle – you don’t design for accessibility because you don’t think you have any learners with disabilities, but you don’t have learners with disabilities because you aren’t including them. 

Rather than making assumptions about individuals with disabilities, we can adopt a different mindset. Start by asking, “How can I remove as many barriers as possible to a person learning this?” This shift in perspective can lead to curiosity about people and how they learn, which will improve your design practice. By changing our framing, we open possibilities of what work could look like, helping us create better solutions that are beneficial to everyone. 

By giving a stage to accessibility, I believe the Canadian eLearning Conference sets an example for the industry in how we can create a more inclusive future. So many modalities and technologies exist for teaching people. If we aren’t talking about accessibility in conjunction with those topics, we’re missing opportunities to reach a wider range of people and deliver a great experience. 

The Canadian eLearning Conference is held in Toronto, Canada on June 13^th and 14^th. In addition to our sponsorship of the Accessibility and Inclusion Stream, I’ll be facilitating a session, Design for All: Essential Accessibility Practices for Instructional Designers, on June 14th from 9:45-10:45 AM. I want session attendees to feel confident about creating accessible content and leave with practical steps they can implement right away.   

To learn more about the conference and register, visit the session information page on the Canadian eLearning Conference website.

1. United Nations, “Human Rights,” accessed June 4, 2024, https://www.un.org/en/global-issues/human-rights#:~:text=Human%20rights%20are%20rights%20inherent,and%20education%2C%20and%20many%20more. ︎