AI & Tech Training

👋 Welcome to TryTami’s weekly newsletter. Each week, we share insights on AI and tech training.

TryTami’s training operations platform lets you scale live, instructor-led training without the operational overhead. Request a demo to learn more:

Most companies invest heavily in training, but still struggle to build real skills. Today I’m sitting down with Kelby Zorgdrager, the founder of TryTami.

Kelby has spent decades building and scaling training companies, including DevelopIntelligence, which was acquired by Pluralsight in 2020.

Today we’re talking about what he learned the first time around, what’s still broken in how companies approach skill development, and why those lessons led him to start TryTami.

In this interview, we’ll cover:

Continue reading for the full transcript or watch the live interview on YouTube.

Dave: Tell me about the training company you started and sold to Pluralsight—Develop Intelligence.

Kelby: We started as a supplier to other training companies. In the early years, we supported companies like Oracle, Skillsoft, and Global Knowledge. Our expertise was Java and open-source technologies, and those companies had a big instructor gap in those areas—so we filled it.

Over time, we decided to go direct to enterprise customers. The main reason was that enterprise customers wanted training programs customized to their engineering teams. The big training companies didn’t really have the muscle to build customized programs. So from about 2010 until we sold, we focused on delivering highly customized, engineering-led training for engineers.

Dave: Customized training was a big focus—what problem were you solving back then, and why did it matter to you?

Kelby: I started my career teaching developers right out of college, and I quickly learned most training programs were very academic. They focused on features and functionality—not how to solve real-world problems.

So when I started Develop Intelligence, the goal was to take real-world problems, walk students through why those problems exist, and how to solve them in real time. It was hands-on and project-oriented—meeting the learner where they were and giving them practical skills they could use immediately.

And honestly, if you jump forward to today: if you go buy something like AWS training from Amazon, it’s still pretty academic and off-the-shelf. It’s not customized to the team. The industry still hasn’t addressed the larger issue: learning outcomes.

Large training companies still haven’t customized programs to deliver outcomes that increase productivity.

Dave: What were the biggest lessons you learned from building and selling Develop Intelligence?

Kelby: A couple things. First, because we were a boutique, high-touch, highly customized learning partner, it was hard to scale operations as we scaled training programs.

For example, the year we sold, we trained about 60,000 engineers live, in person, worldwide. To do that, we needed a large back-office team to support it. Roughly 60–70% of our staff was operational—program managers, project managers, training coordinators. That type of business is hard to scale.

Second: we learned quickly that engineers prefer to learn from other engineers—not professional instructors. That required us to build a large bench of what we called “practitioners”—experts we could subcontract on demand. When we sold, we had about 250 expert practitioners in our network. It takes a lot of work to find instructors with real-world skills who can also teach.

Dave: So the two big challenges were scaling operations and building a bench of vetted practitioners who engineers actually want to learn from.

Kelby: Exactly. Most training companies use the same independent instructors across multiple vendors. The challenge is many of those folks are “professional instructors”—they might teach 40 weeks a year, but they’re not actively writing code. We had to find people who were actually practicing software engineers and have them teach. We had to build a different bench.

Dave: After selling Develop Intelligence, what problems in training or skill development kept showing up and ultimately pushed you to start TryTami?

Kelby: Three things kept recurring.

First: customers wanted live learning in a more on-demand way. Most training companies take 6–12 weeks to organize a class. But clients often have a narrow window—they need training in one to two weeks, and training companies can’t support that.

Second: organizations wanted to move from 3–5 day classes to more short-form live learning—like a 4-hour session. The demand is there, but it’s hard to manage instructor time and logistics profitably for short sessions.

Third: the quality of the instructor bench has gone down, while salaries for highly credentialed people have gone up—so finding strong talent has gotten harder.

The broader trend is clear: engineers still prefer to learn live from an expert. You can see consolidation in the market—Coursera and Udemy, Pluralsight and A Cloud Guru—but engagement continues to drop off. Over the next few years, I think we’ll move toward live expert learning in micro, short-form sessions.

And sure, you can get far on your own with self-learning—using ChatGPT or the internet—but eventually you need an expert you can ask: “How do I deal with this situation?” That’s where e-learning falls down, and it’s also where traditional 3–5 day training falls down.

Dave: What did you get wrong the first time with Develop Intelligence that you were determined to do differently the next?

Kelby: We started investing in operational infrastructure too late. Early on, we ran the business on spreadsheets to manage instructors and deliveries. Then we moved to project management software like Monday.com—which was horrible for this use case. Not until around 2015 did we start building our own operational platform.

Because we were late, when we landed big enterprise customers, we had to add more staff than the business could justify. I think we could’ve scaled faster and more profitably if we had operations software earlier.

We had multiple systems—two CRMs, project management, accounting—everything scattered. Managing live training is kind of like putting on a wedding: there are a lot of details before and after. Just managing the logistics steps is time consuming.

Dave: I got married about five years ago—so I can relate.

Dave: What are companies doing where they think they’re doing training well, but they’re not?

Kelby: On the outcomes side, organizations are still way too dependent on NPS to determine the quality of a learning experience.

At the end of the day, NPS functions like a Yelp review. Some people like the barbecue restaurant down the street, some people don’t. It doesn’t tell you whether training moved the needle on the job.

Another thing I see is what I call the “peanut butter spreading approach.” Because organizing live custom learning is hard and expensive, companies buy 1,000-seat licenses for Pluralsight or LinkedIn Learning. From an HR perspective it looks like they checked the box—training is available. But from an outcomes perspective, utilization might be four hours per engineer annually, so people aren’t actually learning.

And last: many L&D professionals become vendor managers instead of learning strategists. They juggle multiple vendors and logistics, and they don’t have time to think strategically about outcomes.

Dave: That leads us to TryTami. How is TryTami taking a different approach?

Kelby: We’re trying to solve three things:

We’re building a platform that gives an L&D leader or training operations manager a single source of record to manage the logistics of all their deliveries.

The platform also includes an instructor marketplace. Organizations can bring their internal instructors or their own network, and they can also access vetted instructors from our marketplace.

And on the strategy side, we’re building what we call L&D AI agents. They can do things like job task analysis to identify skill gaps, read a PowerPoint and help convert it into a course, and help with registrations and evaluations. The goal is to reduce the effort required to deploy customized live learning at scale.

Dave: It sounds like TryTami solves the operational challenges you experienced—plus the instructor bench problem.

Kelby: Exactly. If something like this existed when I was running Develop Intelligence, we probably could have reduced operational staff from about 70% of the workforce to 30%. That cost savings would’ve allowed us to hire more qualified instructors and deliver better experiences.

Dave: Who is TryTami for?

Kelby: Two groups.

First: training companies trying to scale training operations and revenue—to achieve exponential growth instead of linear growth. If you’re a $3M–$10M training company and most of your revenue comes from live learning, you’re probably managing delivery on spreadsheets. TryTami operationalizes the business.

Second: large organizations that run internal learning at scale. One former Develop Intelligence customer had 40,000 engineers and ran live learning for them on spreadsheets. Same problem—just inside the enterprise.

So we’re building a platform usable by both training companies and internal L&D teams running large numbers of instructor-led courses.

Dave: What’s the current state of the product today?

Kelby: We’ve been developing the product for about a year. We’re planning our first public launch toward the end of January. We already have a couple beta customers—one training company and one internal L&D team. The platform is working; we’re just tying up bugs and loose ends ahead of launch.

Dave: What are you most excited about on your roadmap for 2026?

Kelby: A few things. One is expanding the L&D agents. Instructional designers spend a lot of time turning a PowerPoint or course description into an effective course. Our L&D agent can ingest a deck, analyze it, and recommend improvements for the classroom. We’re building more tools like that.

Second: helping training organizations scale with a true single source for operations—calendar integration, payments, analytics—tools that make the platform more robust and useful.

Dave: Last question. If TryTami succeeds in its mission, what will be different in the world in five years?

Kelby: My hope is TryTami facilitates near real-time connection between learners and experts. An engineering team could describe the problem they’re dealing with, and through AI matching, within minutes they could book an expert that afternoon and get the help they need.

If a team is stuck on a bug for a sprint, that slows the velocity of the entire organization. If they can book a “phone-a-friend” expert solution in near real time, they get unstuck and keep shipping. The outcome is simple: velocity.

If you can make the “911 call” on TryTami and connect with an expert in three minutes, it fundamentally changes how teams work.

Dave: And that’s not just engineering—sales teams could do the same with objection handling or tough contract scenarios.

Kelby: Exactly. It changes behavior and performance.

Dave: How can people get started with TryTami?

Kelby: Go to TryTami.com and request a demo. Fill out the basic info, and someone from my team will reach out, set up your account, and walk you through it.

Dave: Awesome. Thanks again, Kelby—excited about what you’re building and the problems TryTami is solving.

Kelby: Thanks, Dave.

Request a demo of TryTami below:

Until next Tuesday,
Kelby, Dean, & Dave