The One Mistake Business Leaders Are Making Right Now

Thirty-year tech veteran Dave Slutzkin believes the reason companies are not seeing big returns on their LLM investments is a misunderstanding of how people work, not AI.

AI is here. Here on every desktop, in every pocket, in the tools your team uses every day. For some, it’s a marvel of mass productivity, while for others AI-first mandates can feel a bit like a prisoner’s dilemma. If they learn AI, are they training their own replacement? If they don’t, will they be left behind?

According to Employment Hero’s 2025 Annual Jobs Report, a quarter of all workers believe AI is already eroding job opportunities, with nearly half (45 per cent) lacking confidence they could find a new role within three months.

So how do you lead your people through this messy middle? Dave Slutzkin, a serial startup founder and pragmatist of the AI revolution, says, it comes down to time, tools and leadership. “I think it’s important to support people through it, maybe try to instill some curiosity in them, give them the time and access to the tools they need to learn, that can give them a bit of resilience around it.”

No newbie to technological change management, Slutzkin has been building software for nearly 30 years. He co-founded Steppen, a clever health and fitness app designed for Gen Z, before its sale in 2023. Today he’s the founder of Cadence, an AI startup focused on improving communication in tech teams, using AI to curate the daily firehose of information, and a deep thinker when it comes to keeping humans in the machine. 

The path to productivity is a process rewrite

According to Slutzkin, one of the biggest mistakes leaders are making is treating AI like just another software update. This isn’t about just giving everyone a ChatGPT license and hoping for the best. It’s about a fundamental workflow redesign. “If you don’t change the way you work, AI gives you a marginal benefit,” he explains. “But changing the way you work unlocks potentially huge gains.”

Rewiring his thinking to get the most out of AI is something that even he grapples with, “If I try to work the same way that I did 10 or 15 or 20 years ago, but just slot AI into places in that process, it doesn’t actually work that well,” he says.

“I’ve learned that I have to take a step back and rethink the process a little bit. Practically that means I’m writing more documentation up front to guide the AI better, then doing a lot more user acceptance testing on what comes out of it.”

Slutzkin explains that framing AI-as-workflow-redesign is a path to getting time back, not just creating faster busywork. 

Meeting your team where they are

For many employees, this transition isn’t just about learning new tools but can feel like an existential threat. This is especially true for experienced professionals who have built their entire identity around a specific skill.

“If I was a really good writer, or a really good coder or a really good artist,” Slutzkin muses, “and now AI is able to do some of those things in my place I’m obviously going to be thinking, like, who am I now? What’s my job going to be in a year or two or three or five? There are emotional aspects to this that are quite challenging for people.”

It makes sense that pushing AI adoption without acknowledging these identity crises will be met with resistance. Slutzkin points out, “It’s hard to know how to balance that as leadership,” he says. The solution is to reframe the mission.

“Make it clear to them that, yes it seems that this is where the world is going. We’ve got to get you good at this either for this role or your next one. You’re going to need this.” By positioning AI fluency as a career skill you’re helping make them indispensable for the future. 

“I think there are fundamental limitations to the current generations. Not to say that it’s not amazing, or that it won’t continue to be amazing, but it’s probably plateauing a bit and we’re now at the stage of learning how to get the best out of it. I suspect we’ve got a bit of time to solve some of those problems.”

“But I don’t think anyone is particularly taking a really balanced view of what is possible because they’re all running as fast as they can to outpace each other.”

Getting your team to AI ready 

As the famed-Shark Tank US investor Mark Cuban said this year, there’s going to be two types of companies in this world: ‘those who are great at AI, and everybody else that they put out of business’.

Dave believes that getting a team ready for what’s next comes down to a few key investments. The first is time, not just software. “One of the companies that I invest in has just given their entire team of 15 people three weeks of doing nothing else but messing around with AI and seeing how they can use it in their context and building little things.”

He acknowledges that is, “a long time to do literally nothing on the business except reply to a few customer support tickets here and there,” but sees it as worthy. Dave says that it’s also worth keeping an eye on the generational gap. “I’m seeing a really clear and interesting divide when it comes to exploring capabilities.”

“I think it might be harder for young people to get started, but there’s also a lot more appetite for it. Whereas I’m certainly seeing a lot of people a generation older who are kind of pushing back against AI because it’s like. “Oh, not another thing I have to learn. I just wanted to work for another 20 years and then retire.”

To combat this major companies like Disney, Deloitte, and American Airlines are using AI mentoring platforms like Chronus and MentorcliQ to create a symbiotic exchange between employees. The platforms’ algorithms match junior talent’s AI fluency with the strategic wisdom of senior colleagues, with valuable knowledge flowing both ways.

What happens next?

Ultimately, Slutzkin is an AI optimist. He doesn’t believe AI is a bubble, but he does see a market correction coming as the costs of running these models force businesses to get smarter about where they deploy them.

“This stuff is really expensive to run. OpenAI and Anthropic are losing money per query, so we know that over the next six or 12 or 24 months consumers are going to have to start to pay more. That means some things that people are using AI for are going to get less viable commercially. It’s going to balance out when it comes to what we can commercially use AI for and what humans are actually still cost effective for.”

Then there’s the quality angle. “At the moment, everyone’s experimenting with what AI can do a really good job of, what it can do an okay job, and what it can only do a mediocre job of,”  says Dave. 

He sees a future with AI acting as a tool to help humans solve bigger problems. “I take a fairly positive view of this stuff. I think the world is going to look different, largely, in good ways for humans.”

“Maybe we will accelerate the progress of scientific discovery so that cancer could become purely a chronic illness, or maybe it will allow us to magically solve some really big challenges with climate change because suddenly we know how to sequester carbon out of the atmosphere? I think we, broadly, will make AI work for us. It might just make our everyday life a little bit nicer.”

Steps for Talking to Your Team About AI

Here are four practical steps you can take to start leading your team through the AI transition.

Invest in Time, Not Just Tech. The most valuable resource you can provide is dedicated, pressure-free time for your team to experiment and learn. A software license is a small expense; giving your team the space to master it is the real investment that pays dividends in loyalty and innovation.

Acknowledge the “Prisoner’s Dilemma”. Start the conversation by validating your team’s fears. Show them you understand their anxiety about AI replacing jobs or leaving them behind. 

Reframe AI as a Career Skill. Position AI fluency as a non-negotiable part of professional development that will make them more valuable, both in their current role and in their career. 

Lead a Fundamental Workflow Redesign. Don’t just hand out AI tools and expect magic. Guide your team in a collaborative effort to rethink processes from the ground up. Encourage new habits, like providing more context upfront to guide AI and performing more human quality control on the output. Acknowledge that not every tool will be a winner.