Podcast 117: Inside an AI-Powered Learning Content Audit

Podcast: Play in new window | Download

Subscribe: Apple Podcasts | Spotify | Amazon Music | Android | Youtube Music | RSS

EPISODE 117: INSIDE AN AI-POWERED LEARNING CONTENT AUDIT

Robert Gadd – President, OnPoint Digital & MetaLark.ai

How accurate and reliable is your LMS content? That question may not keep you up at night. But if you don’t know the answer, it’s worth a closer look. Why?

When you’re managing a growing list of learning priorities, it’s easy to lose sight of data hygiene. But AI applications of all types depend on clean, complete information. Otherwise, the weight of needless content debt can weigh down business results.

How can you get ahead of this issue?

That’s what I’m discussing with today’s guest, Robert Gadd, President of OnPoint Digital and MetaLark.ai. Robert is a long-time learning tech innovator who is leading the way forward with next-level tools to diagnose and resolve content problems.

So, join us as we dig deeper on this episode of  Talented Learning Show…

INSIDE AN AI-POWERED LEARNING CONTENT AUDIT — KEY TAKEAWAYS

• Learning content can be a strategic asset or a serious liability. And as AI-based systems take hold, they only amplify the downside of outdated, inaccurate or incomplete information. That’s why data quality is fast emerging as a critical issue for learning organizations to tackle.

• Tools that create more content more quickly aren’t necessarily a better solution. In fact, this can be counterproductive if you don’t focus on quality. But digital content audit tools can help identify and resolve key issues. This approach is highly beneficial in a variety of enterprise use cases.

• Running a quick “content MRI” scan before migrating to a new LMS or deploying AI agents decreases business and compliance risk, while you improve user experience and learning outcomes. But to maintain the health of your learning ecosystem, it’s wise to treat content auditing as an ongoing process. 

INSIDE AN AI-POWERED LEARNING CONTENT AUDIT — Q&A HIGHLIGHTS

Welcome, Robert. For those who aren’t familiar with OnPoint Digital or MetaLark, why don’t you start with a brief overview?

Sure, John. For I’m a technologist and serial entrepreneur.

And prior to becoming an LMS vendor I was actually an LMS buyer. I thought the system we selected was clunky and could be improved. So, in early 2002, I decided to build an LMS company. But that was a time of tremendous market growth, and by the time we launched, we had hundreds of competitors. So, we needed to clearly differentiate ourselves.

Yeah…

Our solution has always been high-touch, highly customized to meet anyone’s needs. And we tend to stay at the leading edge of learning innovation. So, in the early days, when many organizations relied on a few major LMS companies that didn’t support mobile well, we found our niche by extending our clients’ existing systems. And OnPoint became known as a specialized player that made complex LMS platforms work as mobile learning solutions.

Don’t miss what’s happening at the leading edge of enterprise learning technology! Get our weekly email brief. SUBSCRIBE HERE →

How did that lead to MetaLark.ai?

Interestingly, we’re not content people. But we recognize good content.

The OnPoint team has never actually built learning content for anyone. But we understand how it is built, packaged, delivered, and tracked. That makes it possible for us to do things like run SCORM tracking on a mobile device that isn’t connected to an LMS.

So, as a by-product of our business, we understand content at its molecular level. It’s baked into our DNA.

I see…

Over the last few years, just like vendors in every market, we’ve been deciding where AI fits best in our strategy. And because we know what is inside our clients’ learning content, we think it makes sense to provide a faster, easier, better way to help them understand their metadata and identify skills that align with their content.

So, diving deeply into content and parsing it is the basis for MetaLark.

Interested in MetaLark.ai? Check reviews, demos and more in our independent Learning Systems Directory! SEE THE PROFILE →

This is a great natural evolution, because your team knows all about LMS content bits and bytes. But messy content is a tradition in our industry. So, does AI make content quality more critical now?

Yes, it’s partly how we got here. But I think it’s about balancing data quality and quantity.

You know, the concept of learning in the flow of work has been around for more than 5 years. But now, we all expect the right answer instantly in any system when we enter “help me do X” in a chatbot.

But what if that response comes from training content that is stuck somewhere in a monolithic LMS? Exposing that information at the time of need is a driver. AI just amplifies it.

Also, many learning tech vendors are introducing faster, easier ways to generate more information. Although this is supposed to drive productivity, it can also lead to more bad content.

Right… 

So, multiple factors are driving an increase in content volume, without necessarily addressing its quality.

2026 OFFICE HOURS: If you’re buying a learning system, why not choose with confidence? Join analyst John Leh for this free webinar series just for LMS buyers. SAVE YOUR SEAT →

Content quality has always been a huge challenge. But now, with AI answering questions in real time, it’s exposing serious issues. So, is that what inspired you to position your learning content audit solution as a “digital content MRI”?

Yes. We want people to easily see the value of this solution, so we landed on the “digital MRI” concept.

Think of it as a sophisticated scanning machine that scans and evaluates all of your LMS content in only a few minutes. And then you can pull up information about it at any time, so you can easily make better decisions.

Nice…

This digital MRI construct fits the process, because we look at the information to suss out key points. Then, we explain what we’ve discovered and we share recommendations.

But we’re not the doctors. We’re just the technicians who run this machine and give you relevant data.

The process is based on machine learning. But the AI comes afterwards, rather than before.

You can’t just feed a SCORM course into an LLM an expect it to understand what’s there. First, you need a core understanding of how the content is structured. And for us, that’s machine learning.

Find out how real-world companies achieve more with learning systems that create business value. Get inspired by dozens of success stories in our LMS Case Study Directory. CHECK IT OUT →

How does that work?

Using all our knowledge of how to analyze content and the math behind it, we break it down. Then, we organize that into a vector database to understand where each item is located and its relationship to everything else.

Next, we use AI to throw pieces of that content at a controlled, private model to extract or infer information from it.

Training this model doesn’t require all the data in the world. We just need your data at its molecular level to understand what it means.

And what does “molecular level” mean to you?

We need to break it down. For example, think of a SCORM course.

We have parsing profiles that understand what a SCORM course is. But every vendor with SCORM packaging as part of its authoring tool does it slightly differently. What works for one doesn’t necessarily work for others.

So, as we tear it apart, we look at the manifest. For example, we can figure out if something is built with Articulate. We can go even deeper to determine if it was built with Rise. And even more specifically, we can detect if it is from a version of Rise from last year or four years ago.

Interested in MetaLark.ai? Check reviews, demos and more in our independent Learning Systems Directory! SEE THE PROFILE →

Wow…

How do we know? Because different versions store information and put it in the package differently.

So, we break all of this learning content into little Lego parts. We train our content MRI parsing engine to chop each of these bits down and understand each of those components. And we tag all this information.

Then, we build a larger and larger corpus of information that can be looked at, sorted, searched, and used for discovery. And then from there we use AI to do trend analysis.

For example, how many pieces of content are the same? It’s amazing how often we find hundreds of files or courses with exactly the same name. Which are the newer ones? Which have more information in them? Which are used more often?

So, the analysis starts to bubble up information that turns into action because we can see everything in a normalized way across this corpus.

Cool…

It’s important to get rid of stuff that is replicated, or outdated, or doesn’t make sense, or isn’t doing what it was intended to do. So, people review the audit report and make decisions about all of that information.

2026 OFFICE HOURS: If you’re buying a learning system, why not choose with confidence? Join analyst John Leh for this free webinar series just for LMS buyers. SAVE YOUR SEAT →

And what does that look like? Is it a dashboard with content MRI results that point out actions you can take to clean things up?

Exactly. You can start with something as simple as search and discovery. For instance, we can detect where specific words or images are located, so you can replace, update, or remove unwanted items. That’s simple…

Mmhmm…

But with insights-level steps, we look at content in terms of pattern and similarity. This is where quality comes in. As we break down each piece of content, we use a rubric of six criteria to measure.

But we don’t change or eliminate anything. We just point to items that deserve attention, so anyone reviewing the report will see what they should be aware of as they move through the audit.

Don’t miss what’s happening at the leading edge of enterprise learning technology! Get our weekly email brief. SUBSCRIBE HERE →

Amazing. And does this process let you tag everything with skills and update your library?

Yeah. Skilling is such a big trend right now. It could be part of the LMS, or the talent management platform, or a third-party thing. But our approach helps you understand which skills are reflected in the content…

… For complete answers to this and other questions about what to expect from an AI-driven learning content audit, listen to the full episode on Apple Podcasts, on Spotify, on Amazon, or right here on our site.

Watch This Podcast on YouTube

Learn More About MetaLark.ai

See my “Hot Take” LMS review on YouTube:

For more details about MetaLark, check our Learning Systems Directory → SEE THE FULL PROFILE

Need a Stronger LMS Content Strategy? Let’s Talk

For independent advice you can trust, schedule a free 30-minute consult below with me, John Leh…

*NOTE TO SALESPEOPLE: Want to sell us something? Please contact us via standard channels. Thanks!