Why the Best Workplace Training Feels Like a Conversation
By Michelle Collins, Chief Operating Officer, CodeBaby
Think about the last time you finished a training course. You clicked through the slides, watched the videos, passed the quiz at the end, and then went back to your actual job. Now think about the moment a few weeks later when you ran into the exact situation that training was supposed to prepare you for, and realized you couldn’t quite remember what you were supposed to do.
That gap, the one between learning something and actually being able to use it, is one of the most persistent problems in workplace training. And it isn’t because the training was bad or because people weren’t paying attention. It’s because of how most of us are asked to learn in the first place. As companies look for ways to onboard people faster, hold onto what they teach, and support learning that doesn’t stop after the course ends, character-based conversational AI is starting to look like a real answer.
The Limits of Traditional eLearning
eLearning solved a genuine problem. It made training accessible at scale, so employees could learn anytime and from anywhere instead of waiting for a scheduled session in a conference room. That was a real step forward, and it’s not something I’d want to undo.
But most eLearning is still a one-way experience. Learners click through slides, watch videos, and answer questions that mostly confirm they were awake. What they usually can’t do is ask a follow-up, push on a point that doesn’t make sense, or get guidance shaped to where they actually are. And for most of us, that’s exactly where learning happens. We ask questions, we try things, we get them wrong, and we adjust based on what comes back. A format that skips all of that is always going to leave something on the table.
Learning Through Conversation
Think about the teachers who actually changed how you understood something. They almost never just delivered information. They asked you questions, noticed when you were lost, and adjusted what they were doing based on what you needed in that moment.
Character-based conversational AI brings some of that back into training that would otherwise sit still. Instead of working through fixed content, a learner can talk with an AI tutor that answers questions, explains a concept a second way when the first one didn’t land, offers examples, and walks them through realistic scenarios. It starts to feel less like taking a course and more like sitting with a coach who knows the material cold. People stay engaged that way, and they tend to hold onto what they learn at a deeper level.
Support That Adapts to the Person
Every learner shows up differently. Someone brand new to a topic needs the foundation laid out. Someone with years of experience is usually after a specific answer or help thinking through a real situation in front of them. One person wants the short version, another wants to follow the idea all the way down.
Traditional eLearning has a hard time with any of that. Everyone tends to walk the same path at the same pace, regardless of what they already know or how they learn best. Conversational AI changes the math. Rather than pushing everyone through one fixed sequence, an AI tutor can meet each person where they are, answering a follow-up, offering another example, wandering into a related topic, or circling back to something that didn’t click the first time.
Picture a procurement professional getting ready for a tough supplier negotiation. Instead of digging through modules and manuals hoping to land on the relevant section, they can ask specific questions about negotiation strategy, managing stakeholders, or handling the objections they already expect to hear. The conversation keeps going until they actually feel ready. That’s much closer to how people build knowledge at work, one question at a time, and when learning works that way people gain confidence faster and are far more likely to use what they picked up.
Help at the Moment It Counts
Knowledge fades when it isn’t used. That’s not a character flaw, it’s just how skills work. Organizations pour real money into training, and people still end up searching for answers the moment they hit something unfamiliar.
A conversational AI tutor changes what training even is. It stops being a one-time event and becomes something that stays available long after the formal program ends. Instead of relying on memory alone, an employee can get guidance at the exact moment they need it, whether they’re prepping for a negotiation, walking into a hard conversation with a stakeholder, settling into a new role, or just refreshing a process they haven’t touched in months. Support is one question away, in the moments that actually shape performance.
Practice, and the Comfort of Failing in Private
Some things you simply can’t learn by reading about them. Negotiating with a supplier, managing a difficult stakeholder, handling an upset customer, finding your way through a conversation that could go in several directions. Those take practice.
Conversational AI gives people a place to rehearse those moments before they’re real, and there’s something here I think we undervalue, which is how much easier it is to learn when no one is watching you get it wrong. Practicing a negotiation in front of your manager, or fumbling a sensitive conversation in a room full of colleagues, carries a social cost that has nothing to do with the skill itself. A lot of people hold back in those settings, not because they aren’t trying, but because the stakes suddenly feel personal.
When the practice partner is an AI tutor, that cost mostly disappears. You can stumble through the same scenario five times, ask the question you’d be a little embarrassed to ask a coworker, and try an approach you’re not sure will work, all without anyone forming an opinion about you in the process. Failing in private turns out to be one of the most reliable ways to build the confidence to succeed in public. By the time the real situation arrives, a lot of the mistakes are already behind you.
A Real Example: ADR’s Sophie
One organization already working this way is ADR International, with a virtual tutor named Sophie. For more than forty years, ADR has helped companies build their procurement and sourcing capabilities, and over that time they’ve assembled a deep library of methodologies, frameworks, training content, and hard-won expertise that organizations around the world rely on.
Rather than turning all of that into yet another online course, ADR worked with us at CodeBaby to build Sophie, a GenAI virtual tutor made specifically for procurement professionals and running on our Geppetto character-based platform. Sophie lets learners interact with ADR’s expertise through ordinary conversation. They can ask procurement-specific questions, work through sourcing and negotiation strategy, get guided coaching, reinforce the concepts that matter most, and run realistic practice scenarios. Instead of searching through materials, it’s closer to having an experienced advisor on call whenever you need one.
What I find most useful about that partnership is what it says about adopting AI more broadly. The value was never the technology on its own. It came from pairing the technology with decades of genuine domain expertise. That combination is what produces a learning experience worth having, and it’s something a standalone tool was never going to deliver by itself.
What This Points Toward
Organizations invest enormous effort building expertise, and far too often that expertise stays locked inside documents, training decks, or the heads of a few experienced people who are already stretched thin. Conversational AI offers a way to unlock it, so that knowledge can guide and coach people in real time instead of waiting around to be found. Whether the goal is onboarding, broader workforce training, customer education, healthcare guidance, or professional development, the opportunity is the same: move past static content and toward learning experiences that respond, adapt, and stay available whenever people need them.
I don’t think the future of learning is about replacing human expertise. If anything, it’s the opposite. It’s about making that expertise more reachable, more responsive, and available the moment someone actually needs it. The courses won’t disappear, but they’ll stop being the whole story. What I keep coming back to is that the most valuable learning has always looked a lot like a conversation, and we’re finally building tools that remember that.
