A good metric for measuring the success of your AI rollout is how much people gripe about it. If no one’s pushing back on your AI strategy, that may be a sign you’re not going far enough. So says George Mason University economics professor Tyler Cowen in season nine of the WorkLab podcast, where we explore what it really takes to become a Frontier Firm: a human-led, agent-operated organization that buys intelligence like it’s electricity, puts it to work like an employee, and compounds it like interest. There’s no cheat code to getting there, but these insights, from Harvard Business School professors and consultants and public intellectuals, will jump-start the process.

1. AI reskilling isn’t a skills problem—it’s an identity shock.

Instructing your employees to remake their professions around AI is also a way of telling them to fundamentally change who they are. That’s why an online course or a PowerPoint playbook won’t sufficiently reskill your workforce. Harvard Business School’s Raffaella Sadun says an overhaul of your workflows starts with articulating the purpose and goals of AI adoption: “Start by specifying why is it that you want your people to be reskilled, what’s the business rationale,” she says. “Create some certainty for the organization that [let’s them think], ‘Okay, this is why we’re doing it. There might be a future for me because the organization is going in that direction.’”

2. Using AI isn’t the point.

Dan Diasio, the global consulting AI leader for EY, believes that for companies to successfully transform with AI, they need three things: mindset, skill set, and tool set. Right now, most businesses are hyper focused on the AI tool set, but haven’t contemplated why the tool set is necessary. “It’s kind of like going to the gym if you’re trying to run the New York Marathon,” Diasio says. Running on the treadmill is how you train, but it’s not the point. “Instead, start to think about what the main objective is and where value is going to be created.”

3. Your relationships are your greatest assets.

Judgment and persuasion will matter more than ever, says Cowen, the economist.  AI can supplement or improve on what humans do, but it can’t replicate them outright. “You’ll need to build out your personal network much more. Who can actually vouch for you, recommend you, speak to what you’ve done with them or for them?” As author and AI expert Pascal Bornet points out, genuine creativity, critical thinking, and trust-built relationships are uniquely human characteristics. Ironically, strength in those areas is what makes someone best suited to working with agents: They’re the skills that determine how effectively you can design agents and write prompts.

4. If people complain, that’s good.

“This may sound counterintuitive,” Cowen says, “but the more unhappy people are, the better we’re doing.” Progress is inherently disruptive—it means less predictability, which makes people uncomfortable. The question for leaders isn’t how to eliminate confusion or disorientation. It’s how to recognize when it signals real progress.

Photo credits (clockwise from top left): Tyler Cowen, Pascal Bornet, Dan Diasio, Raffaella Sadun