In conversation with Jay Vergara: from diversity e-learning to AI, with a stop at breakfast in Tokyo
Jay Vergara and I first connected when he was at Rakuten. We had some very early morning calls together, 7am, if memory serves, and I always enjoyed them. He is one of those people who is thoughtful, easy to talk to, and full of ideas.
We later got the chance to meet properly in Tokyo, which I was very glad about. Jay also sent me to Don Quijote for a shopping mission that turned out to be absolutely essential, as I needed presents for my nephews and nieces and had left this rather too late. For that alone, he deserves some sort of public recognition.
What I like about Jay is that, like me, he has come through diversity, inclusion and broader learning work, but is now thinking seriously about AI. That raises an interesting question. Are people like us simply following the market, or are we bringing something useful into a new space?
I thought it would be good to catch up with him for the site and ask where he thinks this next wave is really going.
Jay, let’s start with the important issue: what was that breakfast we had in Tokyo? I remember it fondly, but I have not seen anything like it in Winchester.
The breakfast. Komeda Coffee's morning service. Order a coffee before 11 and the food comes free with it. Toast, egg, the works. An institution. Worth a flight back to Japan just for that.
Also, before we get into work, thank you again for sending me to Don Quijote. Have you any idea how many family birthdays you saved with that recommendation?
Don Quijote. Hah. So glad it landed. Did you find the dried shiitake mushrooms while you were in there? They've literally won awards. Source of fiber, miraculous for the diet, and also worth its own flight back. I'm pretty sure I've saved more family birthdays than I can count.
We first connected when you were at Rakuten, and I still remember those 7am meetings. When you look back, what were the big ideas we were circling around then?
The big ideas at Rakuten. The one that comes back to me is that you can't just transplant Western D&I content into a Japanese organization and expect it to work. We were circling around how to actually customize it for the context. Cookie cutter content was everywhere and most of it was useless in Tokyo.
You once called me the OG of diversity e-learning, which made me laugh. What made you say that, and how do you see that earlier generation of digital learning work now?
Why I called you the OG of diversity e-learning. Because at the time there was no one else I dealt with who had your capability and quality on the D&I side. Maybe even still now. I needed someone who would bring genuinely customized material into Rakuten and you understood that what works for DEIB in the West doesn't always work here. That's why I kept showing up at 7am.
Both of us came through diversity, inclusion and wider L&D work, and now we are both moving into AI. Do you see that as a natural evolution, or do you think some people in our world have simply smelt where the money is?
Natural evolution or smelt the money? Hitting me with the big questions. Some people are probably following the money but what I see is bigger than that. People are moving too slow on actually learning AI while society is sprinting on the doom and gloom side. The fundamental change is that if you don't learn to use AI, not the chatbot stuff but the real building, you'll get left behind in your career and in society. But if you neglect the human skills like leadership you'll get outperformed by AI anyway. Do both and you're invincible.
Put bluntly: when people from L&D move into AI, how do we know whether they are innovators or just passengers on the gravy train?
Innovators or passengers on the gravy train? You asked me to be blunt so I will be. I am tired of L&D "experts" giving opinions about AI who don't know the difference between an MCP connector and a markdown file. I'm tired of yappers who can quote Bloom's taxonomy and L1s and L2s but can't open Claude Code or Codex in a terminal or an IDE and design their own tools. We've been pushing 'human content slop' for years and the slow nature of our business has made us lax about change. AI agents have changed all of that. Pedagogy still matters but if you're not in there building tools at the cutting edge you're just faking it.
Here's a hot take I could talk for hours on: Instructional designers as we know them will go extinct. What survives is engaging human facilitation, because people will get sick of fast AI slop and crave real humans. The future L&D person is an incredible facilitator with domain knowledge who builds their own tools and consults with the business. Buckle up my friend, it's moving fast. I may have just made some enemies with this answer.
Your site is called Lead Human, which is a strong name because it pushes against the usual machine-led nonsense. What does leading humanly mean to you in practice?
What 'leading humanly' means in practice. I've had spirited debates about this with my business partner Matt Gates and at my MBA at GLOBIS in Tokyo. The answer I keep landing on is that leading humanly is the moat AI can't cross. AI is commoditizing every part of our business. Material that used to take weeks is being generated by agents in minutes. Once that's abundant, and it's around the corner, what humans will actually want is the skills part. Psychological safety. Reading the high context disconnect between cultures. Being a good manager. Handling conflict like a human. None of that gets commoditized. That's the moat.
You’re clearly interested in AI, but you also seem wary of hype. What are people getting wrong about AI in learning, leadership and people development at the moment?
What people are getting wrong about AI in learning right now. Most are pattern matching on what they think AI is rather than actually using it. They claim they understand it, try to incorporate it, but they don't have a firm grasp of how disruptive agentic work really is. Agentic AI isn't a chatbot you ask questions to. It's something you give a goal to and watch execute across your stack. That distinction is everything and most L&D leaders haven't even seen it work, let alone built with it.
What do practitioners from diversity and L&D bring to AI that more technical founders or developers often miss?
What D&I and L&D practitioners bring that technical founders miss. The human side. Engineers optimize for capability. We optimize for whether someone actually changes their behavior after the experience. We've spent careers watching beautifully designed programs fail because the human in the room wasn't ready, didn't trust the facilitator, or didn't see themselves in the material. That instinct is rare and most technical founders don't have it.
Are there areas where AI can genuinely improve inclusion, communication or leadership development, and are there areas where it should be kept firmly at arm’s length?
Where AI can help and where it should be kept at arm's length. AI literally can't cross certain lines. At its base it's regression analysis at scale. Even as it consumes more data now, video, audio, sentiment, it doesn't really understand what humans are or how someone reacts given the intangibles. D&I and L&D folks know this maybe better than developers do.
So yes to AI for content scaffolding, role play simulators, personalized pathways, sentiment analysis at scale. No to AI as the actual coach for hard conversations, psychological safety building, or anything where trust is the unit of value. Use AI to free up your humans, not to replace them.
You’ve worked across cultures and spent years in Tokyo. Has that shaped the way you think about communication, leadership and the risks of one-size-fits-all technology?
Tokyo and cross cultural risk. Living here for over a decade has done something I didn't expect. It made me deeply suspicious of any technology that assumes one default. AI today is built largely with Western defaults baked in. Direct communication, individual achievement, low context everything. Drop that into a Japanese organization without rethinking it and you'll watch it fail in a very polite, very Japanese way. The lesson Tokyo gave me is that defaults are everything and they're invisible until they break.
What are the most genuinely useful applications of AI you’re seeing right now, not the flashy demos, but things that organisations can really use?
Most useful applications of AI right now. I could tell you how I vibe coded my own LMS in Claude Code, yes, fully functional!!! Or how I've cut my video editing down to a laughable amount of time compared to before. Ask me about what repos I use and I'll chew your ear off. Or how MCP with Apify lets me get a true sentiment read on topics I couldn't access through Google before.
Honestly though the biggest one for me is simpler. Crossing academic journals using MCP, Consensus, to actually understand how concepts like psychological safety and manager communication connect to each other. That's the unsexy stuff I couldn't do at scale before.
If someone comes to your work for the first time, what do you most want them to understand about what you are building?
What I want people to understand about what I'm building. Honestly I want them to tell me. I don't know yet. I'm collecting some of my own findings and thoughts in this VUCA world about AI and how it's affecting us as humans, and putting it on leadhuman.ai and on YouTube. Some people seem to like it. In less than a month I'm at around 7000 YouTube subscribers. Please like and subscribe (lol).
And finally, where do you think this goes next for you? More consulting, more product thinking, more partnerships, or something else entirely?
Where it goes next. All of the above. With AI we can't even predict next year, let alone next month. I've got brands reaching out asking me to talk to their people about AI and human skills. A couple of months ago that would have made me laugh.
Wrap Up
Maybe I’ll get to Canada for the World Cup, or maybe Jay will make it over here first. Either way, I am quite sure we have more conversations ahead of us, and probably some good collaboration too.