Harvey Singh, founder and CEO of Instancy, shares with host Nichol Bradford how AI agents will revolutionize e-learning and HR. Singh envisions a future where these AI assistants can help automate tasks and personalize learning across organizations—as well as empower employees by reinventing traditional roles.
Harvey Singh, founder and CEO of Instancy, shares with host Nichol Bradford how AI agents will revolutionize e-learning and HR. Singh envisions a future where these AI assistants can help automate tasks and personalize learning across organizations—as well as empower employees by reinventing traditional roles.
Rate/review AI+HI on Apple Podcasts and Spotify.
Nichol Bradford:
Welcome to the AI+HI Project, a SHRM podcast. I'm your host, Nichol Bradford, executive in residence for SHRM. Thanks for joining us. Each week, we sit down with experts to provide strategic insights, actionable strategies, and practical tips to innovate a future that combines artificial intelligence and human resources. This week, we are discussing unleashing the power of AI agents and automation, personalized, immersive, and on-demand learning experiences for the modern workforce in order to create innovative and personalized workplace learning experiences that drive skill development and business growth.
We're going to explore how these intelligent tools automate tasks, customize learning, and deliver data-driven insights, ensuring that your workforce stays ahead in the fast-paced business world. Joining us today, we have Harvey Singh, the CEO and founder of Instancy, to talk about this subject in depth. Harvey has pioneered AI agent learning and learning systems over three decades. He's developed innovative platforms and advising on generative AI and education, and he's contributed to e-learning standards via the ADL Initiative and IMS Global Learning Consortium, including earning the e-Learning Executive of the Year nomination in 2001. So Harvey, thank you so much for joining us.
Harvey Singh:
Thank you. It's a pleasure to be here.
Nichol Bradford:
So AI agents-
Harvey Singh:
Yes.
Nichol Bradford:
... and HR, what is this all about?
Harvey Singh:
Yes. Well, AI agents is actually something that we can think of is technology built on top of large language models. So, as you know, the buzzword generative AI came onto surface like maybe less than two years ago when became more of a phenomena that everybody now understands as ChatGPT or many other similar technologies that have come about.
So generative AI allows you to be able to generate content in variety of formats, whether it's text, video, audio, and so on. AI agents are the next wave of AI technologies that built on top of the large language model but allows you to actually perform useful tasks and help you automate things, and it allows us to harness the capability of generative AI.
But the tasks could be, for example, booking a meeting for you, or it may be transcribing some text, or it might be to look up some information from your HR system and then be able to operate on that data and maybe send out a personalized message to one of the employees. So we are able to actually use large language models to understand data coming out of these business applications, and we're able to actually automate the tasks and make them a lot more useful in our business enterprise kind of environments.
Nichol Bradford:
And specifically for e-learning-
Harvey Singh:
Yes.
Nichol Bradford:
... how will AI agents help us learn more and better at work?
Harvey Singh:
So I think of AI agents as a powerful technology that would really be integrated into the entire learning ecosystem. So when we think about talent development, skill development, and if we look at the total infrastructure and the learning ecosystem, that means that how do you curate? How do you research content? How do you create content? How do we empower our subject matter experts who are working in the field, who have the knowledge, and they are applying that knowledge in the workplace?
How can they bring that knowledge back and contribute that knowledge to the overall learning infrastructure? In addition to that, how do we disseminate the learning to the right learner, to the right employee at the right time, as well as how do we apply the AI agents technology to understand individual's job roles and the tasks that they need to perform and proactively give them the right knowledge so that they can actually perform at the right and most optimal level?
Then, the next level of this ecosystem, we can think about providing a conversational interface to our users so that they'll be able to ask questions and they can find the answers they're looking for. And they'll be able to use multimodal input, meaning that these AI agents can understand not only textual questions but they can also understand visual information.
They can also understand the audio input so that our knowledge workers will be able to simply ask questions, and they can use their device to point at a particular image of a workplace situation or a document, or they can point to any picture of a product, and the AI agent can help them get the knowledge about that question and answer that right in time when they need it, and they can get the guidance and assistance to do their task in the right way.
So that's the conversational possibility with the new user interfaces that are going to come forward now. Then, of course, the next level of the learning ecosystem is the ability for us to tap into the data analytics. We are going to track every aspect of the learner's activities, not only their learning activity but also their work activities, and the AI agents will be tapped into that data and be able to extract information and help them perform their jobs.
Nichol Bradford:
So I'm obsessed with AI agents, by the way, and so I was so excited to... that you were coming to the show because I'm just obsessed with this. But I'd love to know what is your why? Why are you interested in this? What came first? You've been in e-learning for a while. Could you talk a little bit about-
Harvey Singh:
Yes.
Nichol Bradford:
... your passion for AI agents and e-learning?
Harvey Singh:
Yeah. My passion for AI agents or AI technologies goes back to my graduate studies at Stanford. When I was doing computer science, I actually was in expert systems knowledge-based technologies in those days when the word e-learning actually hadn't come around. And I had also gotten into a graduate school of education there, and my research project was at Apple, researching how technology will transform learning. So my passion was how learners can collaborate and build knowledge as building blocks and they can share knowledge.
So that was my project at Apple, and that research project led me into building a whole career around learning technologies, and I have been involved with internet technologies. Then, how do we create learning objects, and how do we create repositories of learning? How do we apply learning standards so that we can interoperate learning across the enterprise? And then going on to the mobile technology, for example, when the early days of the iPhone and the smartphone, how can we leverage that to make learning happen right at the place where the knowledge worker really needs it? So I have been looking at all these technologies and the waves that technologies that come along and how do we apply them to workplace productivity and learning.
And obviously, when [inaudible 00:07:58] AI came along, it became evident that it's actually going to help us take this whole paradigm of learning to the next level. Basically, it's helping us tap into the knowledge, help us develop and generative AI to create new knowledge as well faster. And also the idea of personalizing learning because the concept of adaptive learning is something I've been working on for many years, but frankly, the technologies weren't there to make it as personalized and intelligent kind of in adaptive learning. Now we're able to do that with the power of generative AI and AI agents, you see. So that's what I'm really excited about, that we could actually take this vision forward now.
All these things that we as learning professionals have been dreaming about using multimedia and simulations, role plays so that the learning is not just going to be just a presentation, like a slide presented on a screen, but it's actually going to be very interactive where the learners can interact with it just like I'm interacting with you and I'm going to be able to talk to an AI agent just like I talked to any other co-worker, and I'm going to be able to tap into a rich multimedia repository of knowledge, and I'll be able to converse with it in multimedia kind of a style. So, like I said, from a learning standpoint, this is the best time to be actually taking all of our ideas from e-learning and applying this wonderful technology to bring it to fruition.
Nichol Bradford:
What's really... Well, one of the things, what really comes through is your passion for collective knowledge, collective intelligence, and really human potential. And that really is the of what we're looking at also with the AI+HI Project, is how do we leverage this technology to really support the potential of humans and for the organization? One of my questions for you is tell me your thoughts on this. So we're in a stage with generative AI and all of the many other forms of AI that have been there for a while where we're at a place where we're starting to reinvent work.
And the learning part of the organization is such a big part of pixelating these roles and figuring out what can be automated, what can be augmented, and what can be enhanced. And so what I feel like I'm hearing is when you describe the different things that generative AI-empowered agents can help someone do, how do you see the advent of agents reinventing specific roles? How does it fit into the reinvention of work for... What do the agents do for us?
Harvey Singh:
Right. Right. So I think that one of the good ways to think about AI agent is, for example, the terminology that Microsoft has popularized called Copilot. So an AI agent is like a Copilot. It's really, for example, another analogy that I like to use is the Google Maps and Google GPS so that it's actually a Copilot. While you are going from place A to place B, it kind of assists you and gives you the information you need about the places and how to get to that place and stuff like that.
So it's kind of like a Copilot that will help us do whatever tasks that we are trying to do. If you're in the human resource department, you need to be able to provide a variety of information, look up information about the employees to support them to follow certain protocols and procedures. If you are working in the shop floor or if you're working as a salesperson in your organization and in the sales aspect of your business, then you need AI Assistant to help you understand your products, be able to connect with your customers, to be able to provide the value to the customers at the right point in time.
So imagine an AI agent will be able to connect to my profile as a salesperson in my organization, who my manager is. It also knows what customers I'm currently interacting with because it can tap into the CRM, which has all the data about my customers. It can also look at what stage the customers at, what stage of buying are they at. So by looking at those data points, the AI agent can also actually look up information about the... my client's company and industry, where they're coming from, and what their mind frame is like.
And by using those variety of different data points, it's going to be able to help me as a salesperson perform my job so I can respond and I can communicate with my client at the right time in the right way so that I can build the right relationship and close that sale and increase my revenue. That's the productivity gain that we're going to get. So, for similar roles, if you look at customer service role.
Nichol Bradford:
Do you have some specific examples you can share? Our audience loves cases. So for the other roles, are there... have you seen in the field people that you've worked with where they're using agents already in some of the other examples you'd like to give?
Harvey Singh:
So one of the examples that I'm working with some of the clients right now is that how AI agents can be used to create content for customer education. So a team of individuals who need to build content about their products. Now, as you know, that that knowledge is very diverse. It has different modalities of learning. For example, step-by-step user guides.
You have to produce user guides. It's cumbersome to produce user guides across various software business processes that you have. Now, we can actually use AI agents to help you watch what you're doing, what steps you're taking as an expert, record that to build a user guide automatically, and it's going to capture the screens what you have visited and clicked on.
And it's actually also going to add some visual details and like a arrow to point to a certain part of the screen, what the user is supposed to click on, and it's going to create a user guide, and you can translate that to more than one languages if you like. Now, that would otherwise take, for example, a few hours of work for a writer to understand what the steps are and to produce a user guide. Now you can produce these user guides in a much faster way, maybe up to 70% time-saving.
Nichol Bradford:
Do you have a... That's a great example. And I can also see HR implications for that. Do you have a HR example of agents being used in a process and it totally has changed it or made it easier?
Harvey Singh:
Yes. I have actually developed some agents that will help us create role-plays. So, for example, if we have to prepare a candidate to talk to a manager or a manager needs to talk to one of their reportees, their direct... one of their direct reports, they can actually play... go through a role-play and learn about what's the right way to talk to the person on a challenging situation. So the interesting thing is that, earlier, when we would develop these kind of role plays, it would take a lot of time to script all the details about what would the manager say, then what would the employee respond, and based on their response, maybe there are three options.
And if you pick option one, then you go down this path and so forth. Now, with large language models, we can simply import a scenario. And in fact, if we have some prior case studies or some information about tough situations with employees, we can simply load them in and generate a role play automatically. And that role play would allow them... we can give a tool to our manager saying, "If you need to talk to your employees, please go through these role plays before you talk to them because it helps you prepare and have the right kind of conversations."
Nichol Bradford:
That's great. And now you can do with voice too, so it can be back and forth. Okay, so final question. Five years from now-
Harvey Singh:
Yes.
Nichol Bradford:
... what does HR and agents look like? What's the future five years?
Harvey Singh:
Yes. Well, as you see, that these AI technologies are getting smarter, and smarter and they'll be able to perform complex tasks that's the high level. We know that's the direction. The other thing is that the AI agents work together with other AI agents. So it's not really actually one AI agent. It's actually going to be a number of AI agents. The other thing that we can see very easily is that every person in the organization is going to have their own personalized assistant that's going to record all their tasks what they're doing.
And that serves as a big memory because if I have a knowledge worker, regardless of their role, it could be HR, sales, customer service management role, or what have you, whatever their role is, you imagine a person having a personal assistant, which is knowing exactly what their preferences are, it's knowing what job they have to do, it's knowing their exact job role and also what tasks they're performing on a day-to-day basis. So if I met with a customer, it has that knowledge. Okay.
If I actually need to follow certain protocols that my company has written down as policies or I have to comply with certain things, it has all that knowledge, and it can tap into all of that information to help me do my job. So that's the future vision I can very easily see coming to fruition in three to five years from now, where AI agents are really built into our devices, and we can simply talk to them. It's kind of like I'm having my assistant with me all the time, and it's going to be helping me do my job at all levels. It has all the data about everything that I'm doing.
Nichol Bradford:
That's wonderful. Yeah. My hope and vision three to five years from now for AI agents is that when we look at the reinvented roles, it allows people to spend... instead of spending 10% of their time at the top of their game, it allows them to spend 90% of their time at the top of their game-
Harvey Singh:
Absolutely.
Nichol Bradford:
... doing what they most enjoy-
Harvey Singh:
They could be creative.
Nichol Bradford:
... what they're the best at and the most human tasks. And so that's where I hope it will be. Well, thank you so much for your time today.
Harvey Singh:
Thank you. It's a pleasure to be here, and thanks for wonderful set of questions. I think you're absolutely right that the excitement of AI agents is just getting started, and it's going to be a wonderful ride. Obviously, things are going to change when we have to change with it.
Nichol Bradford:
Yes, yes. So thanks for joining the conversation, and we'll catch you next time on the AI+HI Project.