How AI Can Make Work More Human w/ Juan Naranjo

AI is already taking on tasks typically completed by people. It’s answering contact center calls, taking drive-thru orders and writing marketing copy. Companies are betting on AI to reduce costs, alleviate staffing challenges and improve efficiency. But where does this leave workers? Is AI just going to replace people, or can this technology actually make work more human? 

Join Juan Naranjo and JD Dillon as they share ideas for leveraging AI to improve the employee experience. Juan applies AI in his work as Senior Manager of Corporate Design and Development in the Canadian telecommunications industry and JD’s spent 10 years leveraging AI as an L&D professional and technologist. Together, they’ll show you how AI can make your workplace more efficient, creative and human. 

In The Know is brought to you by Axonify, the proven frontline enablement solution that gives employees everything they need to learn, connect and get things done. With an industry-leading 83% engagement rate, Axonify is used by companies to deliver next-level CX, higher sales, improved workplace safety and lower turnover. To learn more about how Axonify enables over 3.5 million frontline workers in 160-plus countries, in over 250 companies including Lowe’s, Kroger, Walmart and Citizens Bank, visit axonify.com.

Transcript

JD Dillon (00:12):

Hello, friends. It's great to see you. Welcome to the 35th episode of In The Know, your 25-minute deep dive Into the modern employee experience and what we can do to make it better. I'm JD from Axonify, and today's show, it's all about AI. And I know what you're thinking, aren't there enough podcasts and articles and webinars about AI right now? Plus, didn't we already do an episode about AI? Yes. Yes, we did. Episode 23 was titled The Robots Are Here. And if you weren't able to tune in, we actually used AI to generate the guest, who I then interviewed about the opportunities and risks associated with applying AI in the workplace. So if you didn't see it, you should definitely check out that episode on the Axonify YouTube channel. But here's the thing, that was like six months ago. That's a really long time right now when it comes to the pace of innovation around generative AI.

(01:07):

And the thing is that talking about AI is not just the trendy thing to do. It's something we need to keep discussing because it's going to influence the way all of us do our jobs every day. And right now, there are lots of people who are concerned that AI may make work less human, more generic, more automated and even less creative. But our guest today, he sees things a little bit differently. We're joined by Juan, a next-generation learning professional who applies AI in his work with a leading Canadian telecommunications company. Juan and I are going to dig into the practical side of AI-enabled work, including what tasks people should continue to do, how AI can help people be more creative and what we can all be doing right now to prepare for the AI-enabled workplace. But before we welcome our ITK guest, I thought I'd share a few new examples of how I'm applying AI within my work.

(01:58):

You see, I write a newsletter every week called Ecosystem, and it's where I expand on themes for my book, the Modern Learning Ecosystem and the August 6th edition was aptly titled Gains. And I explained how I used AI to save myself 11 hours on the job in one week. So I thought I'd give you a couple of those examples. If you want to read the full post, it's available at LearnGeek.substack.com. But first, I want to talk about how I used AI to generate the thumbnail for a YouTube video. And as any YouTuber will tell you, the thumbnail makes the video. I mean, have you seen this thumbnail from the episode that we did with Kara North? I don't even have to tell you what that episode's about. You just want to see it, right? And I'm actually using AI already to generate about 70% of the images I use in media nowadays.

(02:44):

And in this case, I was producing a video to answer all of the Q&A from a webinar that I did about AI and learning and development. But I needed that eye-catching thumbnail before I could share the video on the Learn Geek YouTube channel. So what I did was I jumped over to Bing Image Creator, and I entered the prompt, “a variety of robots staring into a white circle”. Five seconds later, I had this image. I then photoshopped myself into that circle space, added a few filters, a little bit of text, and voila, attention-grabbing thumbnail, who wouldn't want to see what was going on in that video, right? I know. I also use Gen AI to summarise the content that I share in Slack every day within my team. Now, I share a lot of articles from across the marketplace, videos and research reports. And the truth is, people just don't have time to consume all of that content.

(03:33):

So what we did was we applied the same generative AI we used to automatically build assessment questions, to automatically summarize content within Slack. So, for example, here's Red Thread Research's brand-new 30-page research report into the learning technology marketplace, everyone should check it out. And here's the AI-generated summary of that report. What the bot did was it distilled down the key points from the document, from an article, from a video, into a simple bulleted list. Now, it doesn't include all of the detail. What it does is it gives you enough context to understand, is this something that's relevant to me? And should I consume the entire piece of content, which everyone should in the case of the Red Thread Research report. Given how much content I curate and how much pops up throughout the course of the week, this just wouldn't be possible before generative AI.

(04:19):

And the last one I want to talk about is audio, specifically voiceover. Voiceover is one of those tasks that is very quickly being overtaken by AI, because a lot of times, you just need a voice, not necessarily a performance. And we've actually been using AI-generated voices here on ITK for several months. I mean, remember that episode with Kara North that had the really cool thumbnail, that show included a narrator that was entirely synthetic, and in the past, it would've taken hours of scheduling, recording editing to create those voiceover assets. And in this case, it just took minutes using 11 labs to generate Victoria. Isn't that right, Victoria?

Victoria (04:55):

That's right, JD. And thanks for having me on your show.

JD Dillon (04:58):

Thank you, Victoria. And that's just a few quick examples of how I'm using AI to make myself more productive in my own workflow. So how are you using AI as part of your job right now? Let us know in the LinkedIn chat, and share some of your best examples. But now let's welcome our very human ITK guest, Juan Naranjo. Juan works in the Canadian telecommunications industry, leading a team focused on next-gen learning. His work blends the worlds of learning science and technology to advance the boundaries of what is possible in corporate education. His team has successfully deployed a proof of concept for a virtual learning assistant and immersive VR course for field technicians and the integration of the company's LXP into the employee's native tools. Juan has combined his knowledge of economics and management with the world of adult education to produce innovative solutions that allow organizations to assess the ROI of their learning function. Juan, you're In The Know.

Juan Naranjo (05:48):

Well, thank you so much for having me here and with that introduction, I want to talk to the guy too.

JD Dillon (05:54):

<Laugh>, very exciting, very exciting indeed. Thanks so much for joining us today. So let's start off right, let's get right to it with a question that is likely shared by a lot of people who are watching us right now when it comes to the impact of AI in the workplace. So I'm curious, Juan, are you afraid that AI is going to take your job?

Juan Naranjo (06:13):

In the short term? No, I think there's a lot to the natural intelligence part of the equation. We have the artificial intelligence part and the natural intelligence part of the equation. There's a lot that we can provide and help to improve learning. So I don't know if I'm being over-optimistic, but I don't think our jobs are at risk in the short term. Now in the long term, a different story and that's a bigger societal conversation about what we're going to do when these systems get so powerful that it will be problematic for humans to keep up with them.

JD Dillon (06:48):

Very true. So from my perspective, one of the best ways to avoid that concern around AI taking your job is to take a close look at your own job and see how you can evolve it yourself. And that means assessing the workflow to find out what tasks can be more effectively done by a machine and as well as tasks that just don't have to be done anymore when it comes to working within an AI-powered workplace. So from your perspective, what kinds of tasks should be delegated to AI, especially within Learning and Development, and then what kinds of tasks should people continue to handle, whether it's using technology or independently on our own?

Juan Naranjo (07:26):

Totally. I think AI is making us super designers in the current world. We have superpowers now. We can do things in minutes that before will take days or hours. So I think anything that will help you advance faster and see more ideas and extend the information available to solve any problem will be a good help. Things are simple helps that I have used with AI at the personal level. It's good for providing ideas, as you said. Now, 99% of the ideas are not that original or that great, but the 1% that is good is good. And then you also know that there are 99% are things that… don't bother with them. Everyone else is thinking about it. So it's a good way of gauging how to evolve and how to create, for example, a good workshop, right?

(08:19):

Because you can see, okay, suggest a structure for the workshop. And then you go through it. And some of them are honestly lame, but there are a few good ideas. Another good use of AI is, okay, help me redraft this piece of text in a way that is more straightforward, will save words, in a more direct style et cetera, et cetera. That's perfect. That's wonderful. Normally that's a good job. Not always, but again, that's why we need the human still to check it out and say, okay, let's use common sense for the solution. Another good way of using AI as just a direct user, not a Learning and Development department, is you can get research done. That has to be done carefully. I had one kind of research task that I assigned to AI about what are the latest and the most important scientists in the last 10 years that have produced papers about effective learning techniques.

(09:22):

And actually it came up with a very good list. But at the same time, one week later, someone in my team decided to ask AI about very good studies or surveys that prove the return on investment of certain L&D interventions. And then when this person sent me the article from Bing in this case, I started checking the details and all the references were made up <laugh>, right? So it can go terribly wrong. You have to know how to use it. So that's at the personal level, at the organizational level and the L&D department, there are a lot of things you can do. I don't know if you want me to get into those ones now.

JD Dillon (10:01):

Give us one example.

Juan Naranjo (10:03):

Perfect. I think you mentioned one of the most powerful ones: a script or text-to-voice solution. It allows you to create voice overs for all your web-based training assets, which we couldn't dream about before because it was extremely expensive, like making voice overs for our WBTs before, if we wanted to do the whole thing, was more than half a million dollars. Now it costs us just a couple thousand just to do the whole thing. For videos, it's great, you put in the script, you get the script in the voiceover for whatever video you created, and you also can get the French-Canadian version of it. And then you can edit and you can redo things. So it's very powerful in that sense. It saves a lot of money in studio time. In hiring vendors who are specialized in voiceover, et cetera, et cetera. That's one example. I have a few others.

JD Dillon (11:00):

Yeah, and I think, I think what you really nailed in those examples is it's important to not think of AI as one thing, or even as generative AI as one thing because it's level of capability and quality and a variety of different tasks varies just like a person does, right? You don't generally work with a person who's just amazing at literally everything that they could possibly do at work, right? Someone's an amazing visual designer, but maybe they're not the greatest writer or researcher. Someone's a great researcher, but they can't Photoshop. That's just not what they do. So I look at it similarly to say how does it function within each individual task? And as a result, how can I leverage it for those purposes? And that's why I like to share those examples from the top of the show. I'm using similar brands of technology, but in very pointed ways to get specific outcomes that are enabling me to do my job differently. So I wanted to see, when it comes to the creative side of working, you mentioned this a little bit, I'm curious, is there a particular example or process that you've worked through that leveraged AI to help you be more creative in the work that you were doing?

Juan Naranjo (12:05):

Oh, yeah, absolutely. That's a great question. Yeah, we have a few examples. One of them is one of the designers on my team was trying to find a very specific type of photograph and went through all the iStock repositories and couldn't find exactly what he needed or wanted. So he got, I think it was Dali to generate an image. And then that image was okay, but then he downloaded the image, modified it, and the basics were right. It was exactly what he was looking for. Again, the execution wasn't quite there, but Photoshop took care of the rest. So meaning, the human <laugh> with Photoshop took care of the rest, and then the image was excellent, perfect. Something that before we could not have had, like, we would've had to hire a photographer and get the exact shot we wanted, and it would've taken forever and would've had to go through, you know companies, you have to go through RFPs and RFIs and things like that is horrible bureaucratics all the time <laugh>.

(13:09):

But with this, you save time there also, in terms of tweaking text, which can be very creative. You're stuck with your style, and then you ask it, okay, could you write it as if you were a person in this type of job or a person with this type of personality? And sometimes comes back with really, really interesting ideas. The other thing that is very interesting too is that the tools we're using, like Creative Cloud, Storyline, et cetera, they're adopting all the AI tools within them. They're in suite. So we have used the capabilities within Creative Cloud Firefly, what it is called. I speak Spanish, I sometimes call it Firefly, and sometimes it is Flyfire.

JD Dillon (13:58):

<Laugh> It's Firefly.

Juan Naranjo (14:00):

So it is, okay, thank you, <laugh>. So basically, you get the images, the things already embedded in the system, and it's just having the openness or leveraging it when you need it. So those are similar examples that we have used for creative work. We have also used it to brainstorm ideas. So in a brainstorming session, we use it as another person attending, and then we put part of the text of the brainstorming and say, ‘Hey, do you have more ideas about this?’ It's very bad at it, <laugh>. It's getting better. The ideas, as I said, are very cheesy, not very innovative, but again, you never know. Sometimes you run into stuff that is interesting. Again, the 10% is very valuable and the 90% it's just average.

JD Dillon (14:52):

Yeah, I think that's why it's important to separate the term Generate and Create, right? Because it's not really creating, per se, it's generating based on source information. So that's why so many of the ideas, you've seen this before, but maybe you haven't seen it lately. So just bouncing that information off of another entity, in this case, can lead to new ideas. And in my case, I'm a big Bing Image Creator fan right now from an image generation perspective. And I did a workshop earlier this week based on my book that was entirely themed Back to the Future. Now did anyone ask for that? No, no one asked for that. But what the AI allowed me to do was to re-generate the slate of images in my presentation to be evocative of the Back to the Future storyline without it being exactly Back to the Future.

(15:43):

So it's just something I've never been able to think about doing before that I could do now with these tools. So I think it's just, it's interesting as you start rethinking your own workflow through different technologies as those technologies evolve. And to do that, we have to know where does the AI fit, but also what tools are available, what do the tools do? How do they work? What are they good at? What are they not so good at? And there are literally hundreds of tools dropping, like their entire business is now wrapped around archiving lists of AI tools. So you can figure out what's out there, what works and what doesn't. As everyone's racing to catch the hype, I'm curious, how do you get past the noise and find AI tools that are really beneficial to your work?

Juan Naranjo (16:26):

Oh, that's super important in terms of prioritization. So we have two things, we have the expansive approach and the condensed approach. The expansive approach is: we have a couple of people in the team who are very focused on capturing and contributing to building up a list of different tools we come across. And then, from the most interesting ones, we pick the top two or three, explore them and sometimes have conversations with the vendors. That's a way we have discovered a couple of interesting things, but the most powerful way to get into AI is to explore with your IT partner. So, for example, in the case of the organization I work for is Microsoft, for others will be Google, for others will be Amazon, whatever, whoever it is, they are already developing many tools.

(17:15):

And just as a short story for this specifically, again, sorry, going back to the same example, the text-to-voice generator. We did a lot of research, and then we went to IT and then they discovered that everything that was being offered out there, well, guess what? Azure Cognitive Services already had the solution or something close enough by a fraction of the price. So that's what we were applying. And that's the same with the large language models. We have bots in our organization that take care of just in time training and they're successful. And then we're adding a layer, an LLM layer (Large Language Model) to the bot. And the way we're doing it is, again, leveraging Azure Cognitive Services in, I'm not saying it's necessarily the best in the world, it's very good actually, but it depends on who is your technology partner because then you're already acting within the ecosystem. You're saving all the cybersecurity processes and et cetera. So something that I will recommend is to make close connections with your technology partners and IT and explore what is already there. And yeah, keep an eye out there for what it is in the market, but leverage what is handy and easy to use that is of immediate access.

JD Dillon (18:32):

That's a great reminder about the fact that it doesn't matter if you're in Learning and Development, you're in HR, you're in Operations, you're not the only person having this conversation, especially right now given how buzzy AI and Generative AI. So instead of having these conversations in silos and potentially replicating work and spending money, you don't need to spend…collaborate, basically. Especially a great example of why it's important to be friends with IT, because they usually have a bunch of cool toys, and you don't even know that they're there and how they could be beneficial to you. And they're great at risk aversion. So if you don't want to make a mistake that would be potentially problematic for your organization, like taking proprietary information and putting it into free software, don't do that. They can also help you avoid those risks as we all try to catch up with how technology is evolving. And then there's something that you mentioned earlier that I want to dive into before we run out of time, which is the idea of equity and fairness. And I think that's a big role within technology and how we can leverage technology is to make sure everyone gets a chance to do their best work. And I think AI really is going to help us elevate equity within the workplace. I'm curious to get your thoughts on how AI can help make the workplace more fair and equitable for everyone.

Juan Naranjo (19:51):

Oh, absolutely. It democratizes access to many resources that were not available before. And you may be, as you said before, short in one skillset, but you have this wonderful idea in your mind, but you're short in a certain skillset, maybe visual design, and you're an excellent writer, or the opposite. Now you have a more kind of equal opportunity to do things because these tools can help you in the areas you're a little bit weaker and areas that you're stronger too. You have something to validate your ideas. So I think that's a great equalizer of the way many tools have been structured because they can be activated by voice in many cases. And other supports also helps people who are neurodiverse or have accessibility needs. So that's an equalizer too. And something else that actually came back to me now that you mentioned this is an equalizer in a way for humans, is that I don't know if you have seen that synthetic neural networks, which are networks that use other neural networks to learn very quickly, crash in quality and become absolutely an utterly useless. So human input is very important. It gives all of us faith and hope because what happens is these networks completely depend on human input. And as soon as that human input kind of decreases or becomes less creative, they just get stuck in the same and same and same ways until eventually, the system implodes. So in terms of equalizing our life and our natural intelligence with artificial intelligence, that's an equalizer too.

JD Dillon (21:41):

It's a lot like the Fast and Furious franchise, as it were. It starts out with a deep, heartfelt story that also involves car racing, and then at some point by the 10th movie, they're in outer space, and you don't know how we got here. So it's a lot of those same diminishing returns. And I think, again, great points about the equalizing nature of technology. And two of the things I've seen recently that I think are just simple but astounding is one: translation. How much faster the translation's getting to the point where it's of the quality that you need to make it a meaningful kind of improvement to your workplace. And I know how many times in my career we've had to make decisions around, which languages will we translate content to or make available to people. And we're gonna do the four most popular or the ones that are required by regulators.

(22:25):

Why can't it be every language that someone might need because while one person may need it, doesn't mean that they're not deserving of it. So that, and then reading levels, another one we talked a lot about where AI has the ability to reinterpret, as you mentioned, information into different styles. So maybe someone that doesn't have substantial reading skills can actually now consume content that was written in a lot of legalese or by people who aren't familiar with different reading preferences. So now, everyone can consume the same content but consume it in a way that makes sense to them. That's some of the stuff I'm most excited about when it comes to the application of AI in different parts of the workplace. And I have one more question for you before we get out of here.

(23:10):

There was a question that came up in a recent panel that I was doing around AI, and someone asked the question of how fast things are moving, should I just wait, right? Should I just wait six months, or 12 months to really get involved in the conversation? Are things gonna be settled down there? Is it a waste of time to dive in right now? So I'm curious based on how you see the technology evolving and how we start our conversation about the impact now versus the impact in the future. How should people be preparing? Should people wait, or should people be diving into this conversation right now?

Juan Naranjo (23:41):

No way. Like you can not wait, you have to be in it, you have to understand it. Even if you're not doing everything you need to be with the flow. You need to understand what's happening again, because the uses you get for AI are very organic, very adapted to your individual organization as to you as a person. So, no, I think AI precisely ('cause it advances very quickly) is a very good idea to be super connected to it, knowing what's happening. Don't overwhelm yourself, don't try to understand everything because it's physically impossible, but stay tuned to the things that may affect your profession that may make the work of your people better. Like to your point, I think I write great, and then you put it through the engine and say, okay, put it at a Ninth Grade level so anyone can understand what I wrote. And you discover, oh my gosh, it could have been way simpler and more straightforward. So that's with everything. Just be playful, try everything and enjoy it. And stay tuned. Don't disconnect yourself. It's the worst mistake you can make.

JD Dillon (24:46):

And I, and I'm with you, I think there's a difference between being distracted by it and a meaningful exploration and kind of awareness. And you mentioned the example of dedicating some capacity of your team to doing the research. And so it's not if you have 15 people on your team, all 15 people don't need to be AI hungry all the time, but how are you having those conversations as part of your work and then finding those immediate gains, like the stuff we talked about today, where right now you can be improving the quality of your work, improving your efficiency, but still keep an eye on how is this changing the workplace, what work needs to be done, how people do their jobs, and as a result in Learning and Development, how we support people doing those jobs. So I think it's a great point why we're gonna continue to talk about AI here on In The Know as we move forward. So Juan, thank you again so much for joining us today. As thanks for your time, we'd like to make a small donation to a charitable organization that you're particularly passionate about. So could you tell us a little bit about the cause that you've selected?

Juan Naranjo (25:46):

Oh, well, so Haven Toronto is a cause that is close to my heart. They take care of homeless people. The homeless crisis in North America is big and well, even though it's good to realize that it is there, it's better if you do something about it instead of just complaining. So if you have any homeless organization in your city, please support it. Donate time, donate money, donate whatever you can because they are not out in the streets because they love it. Some of them may, but it's not the case. So for the majority, the majority are having a very tough time. So donating your time and money to organizations that support them, it's probably one of the best things you can do to have a beautiful city.

JD Dillon (26:25):

Yeah, great cause indeed. And then Juan, how can people follow you, connect with you and learn more about your work?

Juan Naranjo (26:31):

You can reach out to me through LinkedIn. That's the best way to find me and I'll be happy to answer all your questions as soon as I can.

JD Dillon (26:41):

Awesome. Thanks again so much to Juan Naranjo for sharing his insights into the AI-powered workplace. If you had a good time, be sure to subscribe to ITK. Head over to Axonify.com/ITK to sign up for show announcements and reminders. You can also check out the entire ITK collection on the Axonify YouTube channel or listen to In the Know on your favorite podcast app. I'll see you back here in two weeks. Same ITK time. Same ITK channel for our next deep dive into the human side of work. Until then, I've been JD. Now you're In The Know. And always remember to ask yourself important questions like:  What do you call an AI that's obsessed with its Instagram account? Selfie Aware.

I'll see you next time. 

In the Know is produced by Sam Trieu. Visual designed by Mark Anderson. Additional production support by Richia McCutcheon, Andrea Miller, Maliyah Bernard, Tuong La, and Meaghan Kay. The show is written and posted by JD Dillon. ITK is an Axonify production. For more information on how Axonify helps frontline workers learn, connect, and get things done, visit Axonify.com.