How to Keep Up with AI At Work (Without Losing Your Mind)

[MUSIC] Recently, I called on my colleague, Calum Borkers, to talk about AI in the workplace, in our workplace, the Wall Street Journal, and workplaces in general. Do you use AI for work? >> I do. Maybe less so than some folks. Probably it's most helpful during our headline brainstorming session. My editor and I often use it. He, in fact, jokingly refers to it as the intern. And I think that that's often a good way to think about it. [MUSIC] >> I tend to use it mostly like a research assistant, like a way more helpful version of Google. And I've got to some newsroom training sessions about harnessing its potential while avoiding its pitfalls. This is probably where a lot of people are. You're hearing about it at work, you're dabbling a bit. But man, what exactly are we supposed to do? [MUSIC] Calum's been talking to people on all sides of the AI conversation at work, with employees and bosses. His column on the clock is about people's careers and work lives. [MUSIC] How would you describe people are feeling about it? [MUSIC] >> Uneasy, I think, is the word that comes to mind. I mean, I think that, you know, for every worker who sees some time savings, who sees some opportunity, there's another worker I talk to who feels really threatened by AI. And frequently, those feelings co-exist within the same person. [MUSIC] >> So how can you use AI at work in a smart way? What are the tricks and tools that are worth your time? Is it worth your time? [MUSIC] And how do you stay sane, AI frenzy? [MUSIC] From the journal, this is Work Mode. A special series that will air over the next four Sundays, where we take a look at how work is changing. I'm Ryan Knutzen. Let's get to work. [MUSIC] So one of the things is motivating people to learn AI, obviously, is the risk that they could be replaced by AI and the risk of layoffs. >> Sure. >> How real do you think that threat is? >> In the short run, the risk is relatively low for most professions, because most of the tasks that AI can do still require human oversight and monitoring. That very well may change in the coming years. But I look at the so-called AI layoffs of recent months with a degree of skepticism because these are all the companies that are saying, "Oh, we're reducing our workforce because of these AI advancements." >> Well, that's right. I mean, there's a great incentive on the employer's part to look like you're ahead of the curve, that you're some kind of AI genius. It's a whole lot better to tell the public, to tell your investors, "Oh, we've achieved some efficiencies through AI, and we need fewer people." >> Hyper-Nod. AI is here. And it's getting smarter every day. >> What would you say to somebody that is not thinking about AI right now? And just sort of saying, "I don't need to bother with this." >> I guess I'd ask, "Where are you in your career stage?" You know, made if you're five years from retirement, you know, maybe you can get away with it, right? You know, you figure, "Okay, is it really worth investing all this time?" I've got a pretty short runway left. Maybe you can ride it out. You know, if you're 25, you know, my advice would be a little bit different. >> Well, and what would your advice be for somebody who's 25? Gosh, I mean, take a breath, you know, for one thing. I mean, I think it's a very good, good, good. You know, it's exciting, it's a little nerve-wracking, but I just think that none of us makes our best decisions amid a panic. >> Now that we are all calm and not an panic, Kellham says his biggest piece of advice from the outset, is that you need to lean into your human skills, the things AI can't do, and maybe we'll never be able to do. >> So it's communication, it's critical thinking. It's being a reflective person, right? Somebody who's able to say, "This went well or this didn't well," and why? And what do I learn from that to do better the next time? It's being able to be that AI editor or manager who can, you know, not just accept its face value, the AI output, but also evaluate, you know, its quality. I think those are the kinds of skills that will continue to be in demand down the road, more so than any particular job title, which may or may not exist in the future. >> With those human skills, we'll be doing a lot of evaluating of what AI spits out. If you use AI to write an email or analyze a spreadsheet, you still need judgment to decide what to do with that information. >> Callems as the AI future means we'll need to rethink how we define ourselves and work. >> I think a good way to think about it is that we're all managers now, right? You may not have direct report humans coming to you, but if you're outsourcing any of those tasks to a bot, you're now the manager. And so I think that's skill in its own right, right? Is that ability to critically evaluate somebody else's output and judge it to be good, bad, somewhere in the middle, need some tweaking? And so I think that is going to be a big test. I think for a lot of people, because a lot of rank and file workers don't fancy themselves managers, but really, you know, mentally you need to function more like that. >> So for people that work in a white colored job and have not really dabbled with AI much, but feel like they really need to get on the bandwagon, how do you recommend that they start to approach this? Like how should they learn about it and how should they get up to speed and start using it? >> I think one way to start is kind of comparing notes with co-workers, right? I mean, I sort of every workplace, many workplaces kind of have the sort of informal go-to, you know, as somebody who kind of has good tips and tricks for one another. And often that person is a really, really good resource. >> So sort of asked like, "Hey, is anybody using AI? Anybody got any good ideas out there? Let me try to learn from you." >> Exactly. I think that's a really great way to start partly because it's also going to be very role-specific. >> Knowing exactly how AI can be applied to your job can also guide what kind of AI skills you focus on learning. CalM says that sites like Coursera and LinkedIn are full of courses that for most people are too generic to apply to their particular jobs. >> Be really targeted in what you spend some time training on, right? So it's not just a generic AI 101, but it's AI for attorneys. It's AI for accountants. It's a finding something that is going to be really specific to your job. I guess both the job you have now and also where you want to go. >> So there's like a whole range of ways that people can use AI from like writing emails to like building sophisticated agents that do have your job for you. Where should people begin? >> Yeah, and I think you sort of nailed the spectrum. I mean, on the low end, it really is just kind of using it as an assist in thinking of it as the intern, something on the side that works with you all the way up to kind of like a fully autonomous agent or kind of a duplicate of yourself. I think the tricky thing though about the high end of that spectrum is you need to continue training it. >> As you implement AI into your work, there are certain pitfalls to avoid. Like because AI writing can be easy to spot, it can make your co-workers cringe if they notice you're using it. It can hallucinate. So you've got to be sure to check its work. And then there's what it does to your brain. >> For example, I'm thinking of a story I did recently about people who kind of fret about sort of automating all the busy work out of their careers. And so in the short run, yes, it saves me from some data entry, saves me some some menial tasks that aren't all that stimulating, but at the same time some of those folks also told me it's in those sort of quiet repetitive task moments when sometimes they have an epiphany. And so it's sort of this give and take because I think if we're honest with ourselves, we can only spend so much time really in deep thought. I mean, we do need mental breaks. We want a guard against burnout. >> So if workers are able to get better at AI and outsource more of their work to AI, how should they talk to their bosses about it? Because on the one hand, you know, they might want to brag about it and be like, "Look how great I am. I'm the AI superstar." But on the other hand, if they are too open, their bosses might be like, "Great, you've got all this extra time. I'm going to put more projects on your plate." So how do you balance that? >> I mean, what I have found when I talk to people about this is it sort of depends on the industry. So like salespeople, right? It's a good example. I talk to salespeople. They're not shy. Right? They'll tell you everything because they reap the rewards, right? >> Right. >> So if AI makes them better, if they generate more leads, I mean, they have to a good example, right? >> They're trying to sell more stuff. >> They can sell more stuff. >> They're going to make more money and so it's sort of clear. Like, it doesn't matter how. They go for it. >> Exactly. So there's an incentive there. Great. If I can do more, my commission goes up. So those kinds of roles, I think what I'm curious about is whether some of these other white-collar professions, a lot of professional services firms, I mean, like quite literally, it's billable hours, right? It's the hours that are the measure of what you earn. And it's suddenly your weakest fewer hours. Yeah. >> Yeah. What's your incentive to reduce your billable hours? Right? So I think one thing for companies to figure out, or maybe entire industries, is like, how do you structure the incentives for your workforce so that they want to kind of raise their hand and say, like, yep, this is what I'm saying. I'm able to achieve. And Calam says a lot of workers are raising their hands. He calls them AI power users, and they're impressing their bosses and leaving co-workers in the dust. And I think that the biggest thing that distinguishes them is experimentation. It's the lay folks who have just spent some time experimenting. And so what they do is just a lot of trial and error. The power users are very good about tailoring their prompts and then following up, right? Instead of just taking that first draft or being disappointed because it didn't live up to your expectations, you come back and you reprompt. Hey, that was good. Can you make it shorter? Hey, that was okay. Can you tailor it to an executive audience? But the folks that I talked to who were probably most excited were folks who kind of had an entrepreneurial streak. And so they were either using AI to do the tasks of two or three people as they were trying to get a start up off the ground. After the break, one AI power user on how he did it and made a lot of money in the process. So how would you describe it right now your relationship with AI? I'd say AI is it's my coworker. It's my co-founder. It's my best friend. It has enriched my life by, I'd say, speeding up work, opening new opportunities and new doors and giving me access to information to faster and better than I could have ever had before. Michael Ruger loves AI. And he says it's totally changed the way he works. He's a marketing director for a hospitality company in Salt Lake City. When Michael first started experimenting with AI chatbots, he was floored by the potential. It was a surreal moment. It was honestly magical. It was the first time you were engaging with something that felt a bit intelligent. And by the way, I don't have a technical background. I didn't go to school for computer science. I was going to ask you, you don't have any, you were never developer or anything like that and you never have any coding experience? No. I could do like some HTML code, right? I'm sure I'm back in the day. My space page is, you know, a little HTML here. I can make this page green in the font italics. Michael integrated AI into his life by building an array of AI projects using cloud, the AI chatbot from Anthropic. Each project becomes an expert in its particular area because Michael can upload documents to it and give it feedback. And it remembers and gets better than more he uses it. The projects act like coaches that help Michael manage different aspects of his life. He made these coaches for all sorts of things. He built a fitness coach that knows his diet and fitness routines. He's got a financial coach, which has access to his portfolio. And he's also got a career coach. For example, my career coach helped consult me on this conversation. What did it say? What did it advise you to do? I mean, first of all, I said it's a great opportunity and it advised me on. And it said Ryan Knutson is the best podcast host in the world. So. I did. It actually looked you up. The biggest impact has been at work. And his marketing job, Michael oversees a number of hotel and ski resort brands. I have projects set up for each of them. And so if there's work for a hotel or a ski resort, I could take it directly to an AI. You call it like an agent that has been pre-trained on that business and knows the brand and knows the goals and the objectives and knows what we're working on. So he'd help me just like rapidly work faster. Michael started playing around with AI in his old job. When he worked in the marketing department for one of his company's sub-brands, a ski resort in Utah called Snow Basin. For example, every week, Michael had to write a report on how customers felt about their experience at the ski resort. He would take data from surveys, reviews, and general feedback, and he'd break it down into actionable items that would improve the customer experience. It would take me probably three or four hours to kind of go through, understand what is the pulse of the customer. And I would take the data and see kind of what changed week over week. And I would send out an executive summary to our management team. And I would present on it in management meetings. And this is the first thing I built, essentially, an agent for. I called the agent Monti. I see the moose, which the moose is like our de facto mascot at the ski resort. We would make a moose statue on the plaza. Love moose. Yeah. Love moose. So I built Monti. And I fed it. Probably 30 different reports and emails I had written about this. So my hey, here's the data. Here's how I've done it. You've seen it 30 times. I need you to do this for me. And at first, it did it with maybe 80% accuracy. It wasn't that great, but it would get me there. And it gave me a little bit of a head start. And then AI models continued to improve. It got better and better and better. And to the point where last winter, Monti got so good that Monti was coding little video games in the reports and adding little easter eggs to keep people engaged and started turning it into interactive web dashboards. And you give it a goal and it really took it far beyond anything I ever used to do. Michael started to turn to AI to solve other problems at work. And in early 2025, he had a breakthrough. The ski resort, it's called Snow Basin Resort, was largely unknown. It kind of like a hidden gem ski resort. And we had built the brand into being ranked number one by several leading media outlets, like ski magazine, outside magazine, and so forth. But despite that, Michael noticed that when he asked an AI chatbot to recommend ski resorts in the US, Snow Basin didn't really show up, which is a problem when more and more people are turning to AI to build their vacation itineraries. Michael took it upon himself to figure out how to fix that. How can I reverse engineer this? How can I understand how AI makes that decision and start to position my brand in that pathway? In other words, how do I get my brand on the top of these lists? When I asked a chatbot, which is the best ski resort, the one my brand is the one that comes up top? Exactly. And what better way to reverse engineer AI than to ask AI to do it? In his free time, Michael used Claude, his favorite chatbot, to vibe code a piece of software to help him do this. First, he asked AI to create around 50 versions of the question, where should I go ski? And then he built a tool that asked those questions to all the chatbots and exported the answers into a spreadsheet. That turned out, Snow Basin was only showing up in about 3% of responses. Then Michael asked Claude how the chatbots came up with those answers, so we could try to fix it. And then he used his human skills, his expertise as a marketing director, to come up with some solutions, like specific media outreach and incentivizing customer reviews, and updating the Snow Basin website to make it more AI friendly. In the end, his AI mind meld worked. The ski resort went from showing up in 3% of responses to 33%. Michael posted the software he built on LinkedIn, thinking it might be useful beyond his specific company. I put it out on LinkedIn saying, "Hey, I built this thing and I want to test it on some other businesses." Free of charge, I just wanted to get some feedback and see what people thought about it. When I put it out, the response was massive. I had hundreds of brands reaching out. I ended up working with about 50 to test and refine the product before ultimately putting it out to the market in about October. Michael realized that what started out as a tool to help him with his job might actually be a business. So he founded his own startup called Sentium. Sentium offers an AI version of search engine optimization. It helps clients improve how often they show up in AI chatbot answers. And Michael says the company is doing well, with more than 200 brands using it. Michael charges customers a subscription, between $99 and $250 a month. He says he's bringing in around $7,000 a month in revenue. This is a solo business with just me, no payroll, no employees, no developers. This is what this world of AI enables. What it used to take to launch a tech product required going out to venture capitalists and getting funding and hiring great engineers. It's a big risk. So this is kind of a new way of building software. As using AI, you could get there. It's done well. And it pays on my bills. It pays the rent. And it automates itself. And so it has been extremely time intensive to build it. But now that it has built, it essentially runs itself. In a world pre-AI, how many employees do you think that you would have to run a business like this? Oh, a business like this, pre-AI, I think you probably need like 15 to 20 at least. And you would have needed funding to do it. You would have had to go out and hire an expensive engineer. You would have had to convince a venture capitalists of taking a bet on your idea. And it would certainly be in debt right now. But you would be building for this kind of future payoff, right? And so now with AI, you could essentially do it yourself. Aside from his marketing job and his startup, Michael also has a third job. He's a marketing instructor at the University of Utah. He says his students are nervous about what AI will mean for their careers. So what advice do you give your students when they ask you about, "Hey, what did it mean for them?" I say the biggest thing I'm trying to instill in them is a mindset. I tell my students. For the history of marketing, you would graduate from school. You would join the marketing symphony and you would play an instrument in the symphony. Maybe you over here are going to play the drums. Maybe you're going to play the flute. Maybe you're going to play the trumpet. And those are, you know, email, social media, website. You pick a function of the business and you run that function. You're playing a part of the symphony. If you use AI well, you step back and your rule becomes a conductor instead. And you're deploying AI into the instruments and you're conducting the symphony. And let's say this class would have all been a part of one grand symphony and now this class can be 60 different symphonies and they're all conducting their own. And that's, I say the mindset I'm trying to instill in them is it's not just something to help you write emails better. It can be used for far more and you need to think more as someone deploying systems. I mean, I think that that was a really great way to describe AI and sort of where it seems like it's headed. Also in some ways, the fear that's sort of baked into that because it may mean that there are a lot of people that should be conductors that will be able to be conductors because they've got what it takes to conduct an orchestra to make something great out of all these pieces. But there are a lot of people that just want to go do their function, play the saxophone, you know, hit the triangle perhaps if you're not very musical inclined like me. And that those opportunities will be fewer and that everybody wants to be a conductor. No, I think that's true, right? I think, yeah, student graduating right now could go play the trumpet in the instrument in the orchestra and there's jobs for them that exist in that space right now. But future-proofing and fast-forwarding years into the future, I do think the way we do business will probably change pretty dramatically. I think eventually some of these things will become obsolete and you have to adapt into the new way of business. This has been great, Michael, I really appreciate it. Hey, thanks Ryan. Work mode is a series from the journal, a co-production of Spotify and the Wall Street Journal. This episode was produced by Enrique Perez de La Rosa and edited by Catherine Brewer and Laura Morris. I'm Ryan Knutson. By the way, let us know what you think of this episode. We'll send me a message on Instagram @Ryan_Knewtson. We'll be back next Sunday with our next episode in the series. Until then, have an amazing week. I hope you get a promotion.