AI-driven Workforce Transformation: A Conversation with Violet Zilman

CommLab Podcast with Violet Zilman

Sherna Varayath 4:04
Hello listeners. Welcome back to the eLearning Champion Podcast where we explore the strategies, trends, and successes shaping digital learning. Are you ready to discover what AI powered learning and development looks like? Today's episode explores AI transformation in L&D, the need for upscaling and reskilling. For this session, I'm thrilled to introduce Violet Zilman.

Violet brings a unique blend of global experience and academic rigor to AI driven workforce transformation. Having travelled to 28 countries and lived across Europe, Asia and the USA on government assignments, she combines master's degrees in pedagogy, linguistics, and psychology with specialized AI training from Wharton.
Her management expertise shines in leading multicultural, multigenerational remote teams of independent contractors. Violet has developed a proprietary framework for rapidly onboarding team members, with AI tools focusing on delivering projects on time with maximum efficiency while navigating the human side of AI adoption. We are thrilled to have you onboard today, Violet.

Violet 5:21
Thank you so much. I'm thrilled to be here.

Sherna Varayath 5:27
Dear listeners, before we dive in, make sure you're a true eLearning champion by hitting that follow button wherever you're listening in from. Now to get started. So Violet, how do you personally use AI to become your team's MVP, and how can others do the same thing without being tech experts?

Violet 5:47
Well, AI plays a huge role these days because it's not just a desirable tool, it's a necessity. And by having AI at your disposal, you actually can create more time to be available for your humans, and humans can do the same. For example, when I bring people in my team, they are independent contractors and they work through the virtual projects. We do not see each other. It's part of the regulations. And when they come on board, the first few days is very intimidating for anybody who is starting a new position. And AI plays a huge role because I created a custom GPT and sometimes people are intimidated to ask as many questions as they want. And custom GPT provides them that opportunity, because whatever they have in their mind, that custom GPT can answer for them. And I know for sure the answer is validated because this is a custom GPT. It has no information that could be misinterpreted, could be a hallucination or anything else. So this is my source that I collected over the years. And they're valuable, absolutely valuable information that is true, correct and up to date. And by having that source, it eases off people. They understand that there is nothing that makes them uncomfortable.

And for some people, when they're not familiar with AI as much, that breaks them from, oh, I don't need it, I don't know how to learn it into being enthusiastic and comfortable with transitioning into a deeper usage of the AI. So that is one of my main hacks because I think it removes the pressure because we are curious by nature, but asking questions may be perceived as being annoying. And when people have this freedom to ask as many questions as they want, they take out the anxiety out of themselves and they're fine. I use the same technique with the required documentation which is mandatory, but it's 400 pages. Nobody reads 400 pages, so I introduce my team members to simple prompts. And when they see the book with 400 pages that they feel obligated to read and then have access to the same library but with a simple prompt, they can get the information that they need without that effort, it also makes them very happy. And I become everybody's favorite from the beginning.

Sherna Varayath 8:56
Right. I also believe it makes team members more productive as they get what they need quickly so they can use their time for other tasks.

Violet 9:07
Absolutely. Because the settling is a very important part, and educating them through hands on experience is probably the best value that L&D can deliver.

Sherna Varayath 9:19
Yes. Now coming to settling in, what are your thoughts on AI avatars being used in learning?

Violet 9:34
AI avatars are fantastic. Unfortunately, they have been misused. We saw so many bad uses of the avatars. When the company gets bad material to the Avatar and it becomes a bad video with the Avatar. I think digital avatars could be very handy when we are onboarding people. But I know that people are very sensitive to the
hand movements and mouth movements with avatars. It could be like a stopping block, but I also found digital avatars a valuable tool when we are talking about knowledge transfer, those long time working employees who know everything from start to finish. They are not comfortable to be on the camera and share their knowledge in the professional setting. But if you will create their digital avatars, they feel liberated, and the knowledge sharing happens naturally. I think this is one of those fantastic options to ease off pressure and help them transfer knowledge in the best possible way, especially when the company is global. It's really hard to schedule a video production, and it's also very expensive to record someone and to make them into a movie star. But with a digital avatar it's a very simple process. It could be done within a few hours, but that Avatar can serve the company well and the person does not feel a burden to leave work, get in front of the camera, smile, and read the script. Not everybody likes to be on the camera.

This is like one of the probably best applications in my mind. But again, with the digital avatars, we are not outsourcing humans, we are actually helping humans to live their lives and having digital avatars as their representatives in the moments when that is important. They are not substituting humans, they are assisting humans.

Sherna Varayath 12:17
Correct. Assisting. Well, yes, that's a very nice way you framed it. Now, could you give us four more ways how AI avatars can be used in L&D?

Violet 12:34
Absolutely. In L&D, the Avatars could be your coach, that could be someone you can talk to. And you have an option in voice selection, in gender selection, in tone selection. It's a very great way to have it done.
It could be your helpful assistant when you are practicing, for example, your pitch. You can actually have an avatar to go off instead of just a blank screen. That could be your guide.

Avatars are wonderful when they work as the welcoming screen for someone who is just starting with the company. Our HR is using onboarding avatars as well. Again, the extension of the person for the events where the presence is formal, not real. That is the very good use of the avatar.

Sherna Varayath 13:51
Wow. Interesting. So from your experience, Violet, how is AI reshaping learning and performance in real workplaces? Not in theory, but actual change.

Violet 14:06
I think we're living in a very interesting times because everything is being reshaped so fast and at the same time I see the effect of the AI times through 4 generations. Because I work with baby boomers, I work with Gen. X, Gen. Y, and Gen. Z. And it's very interesting how AI affects every single generation, where the baby boomers and some of the late Gen. X’s are still kind of undecisive in which way to go. They consider AI as something they have to learn. They probably don't want to, but they will because they have to. This is the requirement and they're taking it like a chore. With the Gen. Y, where the people take AI as a tool. They are more comfortable in using it, but they're still functioning in combination with personal skills. And AI and with Gen. Z, that is a very interesting observation, because this is a technology driven generation and this is the generation I love. But I also worry about them because they're so used to outsource everything to AI and they are the ones who are giving up their personal skills before they fully develop them. And that's the moment that I'm facing. But with learning and development, it is crucial to address each generation differently in the learning programs. And AI makes it possible because you were not able to customize the learning experience to the point as you can right now because just from the basic onboarding, you can actually mainstream a generation into the route that is working for them. Even the presentation of the main knowledge is different for each generation, like Gen. Z. They are striving on microlearning, but not everything could be presented through microlearning. So you have to be very creative in forming the information that they have to consume. The media that they’re used to, the colours, the voices, the background scenes, they have to be digestible for Gen Z's to actually understand and comprehend the information.

Gen. Ys are more adaptable. Gen. X and baby boomers, they're like we have to do it, we're doing it. So they're like the easiest going. But compared to Gen. Z, Gen. X and baby boomers, they build their knowledge skills in the way that if AI is gone, they still can function.

But with Gen. Z, if AI is gone, they won't be able to function unfortunately, because these skills and critical thinking in analyzing, it takes repetitive activities throughout the time where you actually train your brain to develop the functions. You cannot have critical thinking develop to the fullest when you just graduated school or college. It takes time to develop reasoning, and that's the part where AI is reshaping learning and development because the emphasis becomes on the skills that were not fully developed. So with Gen. Zs, it's more of a soft skills approach in training. With Gen. Y and Gen. X, it's more of a technical approach in training. It's very interesting, but it could be accommodated with AI.

Sherna Varayath 18:47
Wow. OK. Now, where do you personally see the real value of AI at work and where do you think the hype really outweighs the results?

Violet 19:06
Well. The mistaken approach about AI, that it's about automation, like automation for the sake of automation, I don't like it. AI is a very deep and valuable tool that helps human to augment their skills, not to replace them and not to diminish them.

The more I work with AI, the more I love my humans because humans have a very unique approach. And that is the main factor that is playing a crucial part these days because when people build, this internal fear is that it's either too late or too complicated to learn AI and it's very easy to understand why. Because there are so many tools people are using, all those jargon words and whatever. But the main point is you don't need to learn everything.

Not everybody needs AI agent. Not everybody needs to have a full spectrum of AI tools. You can be very productive and deliver amazing results if you learn one or two tools. For example, I definitely recommend Gemini and I do recommend ChatGPT because they do have a very good selection of tools within the model itself. And if you are comfortable with those two, you can do anything you want and be more productive because AI can help you with data analytics. It can help you with documenting your notes. That's my favorite function through meetings, so I just run my AI and it just creates immaculate notes for me, which I can work with.

So things like that, but it's not a substitute. It's in addition to the human brain. And I cannot probably emphasize it enough because you don't need to know every single AI tool. It's impossible, but you need to know a few. You have to know how to use them to the full capacity because it's a fabulous tool. I always give a visual example like you are given a helicopter, but you're using it as a bicycle. You can actually fly high, but people are not understanding the full capacity of those just two tools.

Sherna Varayath 22:19
Wow, that's a nice example. So now, Violet, what are your go to strategies when co-creating content with AI while making sure your human voice stays present?

Violet 22:38
Absolutely. The longer you work in a certain industry, the more you create your experience and that experience is your core. Because the fascination with AI, like I see on LinkedIn and everywhere else, and I can always spot AI generated content because it's the information that is sitting on top and you understand that it has been generated by someone who is not an expert but wants to position himself or herself as an expert.

But when you're creating your content, you understand the core message, what you want to deliver. For example, you are very specific about the outcome you want to get. And when you are outsourcing AI, it’s pretty much not to substitute your thinking but actually expedite certain things that you want to include in the message. For example, when I'm creating my messages, I use my custom GPT. I do not outsource through the language models through the outside of the company GPT. I go to my proprietary. And I understand that when I'm providing a request to find this information, I know that my custom GPT will respond in the way from my source that actually will deliver the results that I want. And it saves a tremendous amount of time. This is actually my recommendation for anybody who is listening. Create your own knowledge base, because knowledge will become very expensive if you will consider the amount of money people will be paying. And the subscription for any AI tool with a good database in the future, it will become more expensive. So whenever you have an opportunity to digitize your knowledge and create your own knowledge base that is based on what you know, that is the base form of existence. Because in that case, you know you're always accurate, efficient and up to date because you constantly work with your knowledge base. And the more information you have in your knowledge base, the more independent you become, regardless of the AI tools out there. So it's all about my perspective and I cannot say it enough to my team members and to anybody. You always have to be prepared for the day when AI is not available, but you have to function. So this is the strategy of using AI. How can you utilize AI in a way that it backs you up now, but when it's gone, you're still able to perform to the best regardless of the technology around you?

I think this the most valuable thing, but building your personal knowledge base with the books, with the notes, with the charts, with everything you know, that is probably the best investment of your time you can allow yourself to do.

Sherna Varayath 26:10
Interesting, yes. Now what values guide your approach to ethical AI use in learning and how can organizations stay more accountable?

Violet 26:27
Some people don't understand that AI is not Google. It's a much deeper and bigger tool, but unfortunately all the mistakes are happening when people mistakenly use AI in the form as they use Google. Because this is a completely different approach and the information is on a completely different depth.

And the harm from that basic search or the use of information that you shouldn't do, but you did just because you didn't know is something that has more severe consequences with AI than with Google. And having said that, AI ethics are very important. You have to operate within the boundaries. You have to understand the security requirements of the company you're working for. You have to understand your personal individual contribution to the AI as a system within the organization.

Every organization has a different structure, but cybersecurity with AI is the most critical part that you have to be aware of and educate yourself on what is allowed and what is not allowed within the company. And learn the rules and make them like a tattoo. It has to be with you all the time. You cannot forget about it even for a second, it's mandatory. I love that every open source with AI is open for public and I carry the same approach with my team members. When I share the information, I’m usually very transparent about how I got the information, where it's coming from and I'm always open to questions. I welcome questions and I create the same culture with my team members. I want them to be open. I want to make sure that they feel psychological safety within the team. So if they have any questions, they should not have any hesitation to ask because I would rather have them ask human first before acting on their own and then creating consequences that are not desirable. And then it will be a delay in fixing the consequences.

Like a simple example with AI, when you're working with a document that you want to scan and you're not following the protocol on security. AI reads any text message that AI sees on the page. We humans see only black on white, AI can see white on white.

And even a document, it was a harmful embedding, it can affect AI on the level of the system, and it can alter the data processing. It's very damaging when people are unaware. It's a very easy way to scan the document prior to working with the document by asking AI to number the lines of the document. And what happens? AI doesn't like to absorb the information from the document, but when you see that you have your lines going but the harmful message written in white on white that you do not see, AI will recognize the number that as well. And you will understand right away that the document is actually a harmful piece and you have to contact your cybersecurity team and address the issue because then you can trace the source and stop the harm before it progresses. Always ask AI to number the lines of the document and you can see hidden messages behind just by using the simple tool. There are many tools, but this is a very easy one because you can visually assess the safety of the document.

Sherna Varayath 31:44
Right. Now how would you recommend someone in the L&D space to build their AI literacy, especially if they're just starting out?

Violet 32:03
I think the hands on approach is the best. You need to experiment with AI in order to understand what the tool is about. Overcoming the hesitation is the probably the biggest part. As soon as you will dive in, you will understand how easy it is. I know some people are uncomfortable but there are lots of courses and they're free. That will teach you the basics of AI. You do not need to know in the beginning the depth of different large language models or whatever. You do not need to overload yourself with the knowledge that is not practical, but when you learn a simple prompt and you will get the result with a simple prompt for something that would have taken you hours to do that is excellent. As an intrinsic motivation, because then you’re open to experiments, but again, people have to zero in on the AI as a tool. It's not a car. I know that people take a very academic approach to learning AI, but I think the opposite is more practical and works better. You have to dive in without fear in the beginning. I will recommend you start experimenting at home, not at work for the safety issues, but you can use AI as your best critic. AI doesn't have to agree with you on everything, but when you want AI to analyze something, you can always ask AI to point out the weaknesses of this document or what could be improved. And when it comes from AI, it's not as intimidating as it is from your boss or your peer, where you could be exposed like you don't know something. But every L&D professional has to understand AI from the perspective that this is a wonderful assistant. But you have to train your assistant. You have to be very specific how you work with the assistant and your assistant will serve you as good as you ask him to. This is the point.

The quality of data in AI equals to the quality of the output from AI. So when you're starting something with AI, you always have to come from your personal expertise to validate the output and then slowly develop this into a multi-level conversation. I love AI for the multimodal approach where it can create a conversation between four or five AI agents talking at the same time. So you can assign a role to each of them, and they can give you a feedback at the same time from different perspectives. That makes collaboration on the documents much deeper, richer, but you are in charge. And that's why when you're starting with AI, start with one tool. Focus on what you know and how you can move in that direction first. Slowly start building the knowledge base and adding to that AI tool as much as you can because that’s going to be your toolbox that you will take with you no matter where you go.

Sherna Varayath 36:47
Interesting. So coming to using AI as a toolbox, how can one use AI to drive a behavioral change and not just deliver information?

Violet 37:01
Yeah, that's one of my favorite topics because every team member has to function, but our ability to function directly relates to our ability to control our own emotions and ourselves. And unfortunately, that is the weakest part of the humans because our thoughts sometimes drive our behavior and the ability to recognize the thinking pattern that is not suits you well, ability to stop yourself right in the middle of something that shouldn't happen, but your emotions are pushing you towards that. That is a great opportunity that will allow you to actually pause, right down to AI.

What are you experiencing right now?

This is not a professional advice, but you can ask for the mindful tools from AI. How can I like remove this uncomfortable feeling? And it's very interesting how AI can guide you through the breathing exercises, how it can take you away from the stress, how it can guide you to separate yourself from the situation, take a moment and then come back. This is priceless because it takes a very good awareness in person, after years of practicing to have the skills. But when I introduced my team members to this simple tool, when you feel like you are in an uncomfortable situation for whatever reason, just go to custom GPT for the mindful exercises and it guides them through and they become hooked. Because it's all about that split moment when humans could make something that they will regret about later versus stopping the cells and that ability is not given. It comes with practice and awareness, but it actually serves as a great tool for people when they are working remotely by themselves. They with their thinking, with their emotions and no one can help them out. And having those simple exercises actually helps them a lot.

That is part of the coaching that could be created with AI, and they're strictly mindful exercises right there for people to experience in a moment of stress, to remove the stress and come back to work. I think it's a great support for the remote workers mostly, because working remotely is for some people is a plus, and for some people, it’s a big challenge. And this ability to have this, it's very valuable because after all, mindfulness changes the behavior. This is wonderful because the people who never experienced that for them, it's like wow, I didn't know about it before.

And now this is something that they go to. And I'm very thankful for my experience travelling that I discovered the power of meditation, how to regulate yourself through breathing through meditation and mindfulness, the simplified Western approach. It's not an in depth tool, it's very easy and harmless, but a very helpful tool that changes the behaviour tremendously.

Sherna Varayath 41:33
Right. So in your view, what's the right mix between an AI driven automation and human connection in development programs?

Violet 41:47
Yes, I love my humans. Humans are the driving force. Animation comes secondary. The reason for that is automation is a wonderful convenience, but at the same time we don't need to give up the processes that we run just because AI can replace it or do it faster.

We have to be in charge of the routine, of the process and of the thinking. And we have to do it for the sake of safety of the company, in the long run, from the moment when AI is gone, we still have to be able to function. That's why the automation has to be on the second burner. But it's also important for the human brain, because one of the most favorite automations we all love is the GPS. We use it all the time. In Japan, they ran a very long video study on the effect of GPS on human brain. They did it from 1984. And the effect of GPS, and this is one of the longest, probably automations that we use too has this. It has the same effect on the brain as Alzheimer's. People who use GPS all the time, actually shrink the parts of the brain that are responsible for navigation.

It's a huge difference between navigating through the screen in the car or navigation when you are aware of the corners, trees, houses, and everything else. And GPS became a habit. But unfortunately in our brain the rule is whatever you don't use it, you lose it.

So the long term effect of automation on our brain, a massive automation that we're experiencing right now, could be very detrimental. That's why we do not need to automate everything. We need to keep doing what we are doing because we want to function as humans and the ability to function as humans we have only when our brain is working.

And meditation is the offset to automation, because during meditation, you restore your brain to the functioning condition. It's very interesting and that's why in Japan again, they are ahead of us in the studies of the human brain in connection to automation, they do develop lots of programs where they incorporate writing with pen and paper. Reading paper books because it's not the old fashioned model. It's practicing sensing the sensations that actually make our brain alive and functioning better. And that's what's very important. So that's why, going back to automation, I truly believe that animation should be 60/40, 60% with humans in charge and 40% being automated. And by 40% automation I mean mandate boring processes that actually humans are not benefiting from, but AI can outsource and do.

This is one of the things that could be outsourced, but I don't believe in outsourcing critical thinking. I don't believe in outsourcing creativity. I do not believe in outsourcing emotional intelligence where AI will come up with the response. I do believe in human qualities and capacities, and they have to be cherished and preserved.

Sherna Varayath 46:27
Wow. Interesting. And this brings us to the end of another insightful episode of the eLearning Champion podcast. We've covered a lot today from behavior change and AI driven automation and technical use of AI. Thank you so much Violet, for sharing all your thoughts with us.

Violet 46:46
My pleasure. Thank you for showing interest in this. This is a very important topic and something that people can benefit from right away.

Sherna Varayath 46:57
Yes, absolutely. And to all our listeners, I hope you're walking away with innovative ideas to transform your L&D with AI. Do take a moment to share this episode with your colleagues and peers. Send us your suggestions on your favorite social media platform. Thank you so much for tuning in to the eLearning Champion podcast. Until next time, take care and happy learning.

Violet 47:22
Thank you.