Most organizations think they have an AI problem. They actually have a learning problem.

That distinction came up early in Candice Faktor's conversation with the Toronto Region Board of Trade during Toronto Tech Week. Candice, Co-CEO at Disco, sat down with host John Warren for an episode of Toronto Talks, and the conversation cut through a lot of noise around AI adoption and AI upskilling in organizations.

Here's what stood out.

The tooling trap in AI training for employees

Organizations keep framing AI adoption as a tooling decision. Which platform? Which model? Which vendor?

Candice was direct about this: "You are completely underestimating what it actually takes to build capability in an organization." Handing everyone access to an AI tool doesn't make a company AI native. It just means everyone has access to tools they don't yet know how to use well.

The real work is the human element: helping people move from fear to fluency, from fixed roles to fluid ones. That's what genuine AI training for employees looks like.

Three kinds of people in every organization

Candice used a framework from Brian Balfour to break down how people respond to AI change. Leadership strategy should differ for each group.

The first group is catalysts: people already energized by what's now possible. They want to move fast and show others what AI-native work actually looks like in your context. Your job with catalysts is to remove friction and let them demonstrate what's possible.

The second group is converts: solid workers who want clear parameters. What's acceptable? What's exceptional? If you define that clearly at the role and workflow level, most of this group will move toward it. They're waiting for direction. This is where AI change management done well pays off most.

The third group is anchors: people stuck in older paradigms. Candice's take here was more nuanced than you might expect. Their caution often points at real risks around compliance, privacy, and judgment. Those are worth hearing out, even if the pace needs to change.

Why playing it safe is the actual risk

There's a version of "being careful about AI" that feels responsible but is quietly dangerous. Candice named it plainly: "If you're not taking risks or experimenting or trying new things, you're not safe."

The asymmetry matters. You don't have to bet the company on AI transformation. You just have to start. Try something. See what works. Learn from what doesn't. Then do it again.

The people and organizations that wait for certainty before acting are the ones that fall behind while everyone else is iterating.

What great AI leadership requires today

Candice laid out three skills that separate effective leaders in this moment from everyone else.

The first is curiosity. The ground is shifting every two weeks. Leaders who delegate their AI understanding to someone else are orienting their entire organization with incomplete information.

The second is empathy. Command-and-control doesn't work when you're asking people to do something risky. People need to see their leaders being open about their own learning, their own uncertainty, and their own experiments.

The third is strategic clarity about where agents fit. Which work should agents do? What does that free humans up for? What does it change about your competitive position? Efficiency gains from AI are table stakes. The more interesting question is what new growth opportunities open up when the routine work is handled.

AI upskilling as a continuous learning culture

One thread ran through the whole conversation: the organizations that will thrive aren't the ones with the best tools. They're the ones that build continuous AI upskilling into the culture itself.

Candice put it plainly: "AI can replace people who aren't interested in building continuous learning."

That's the real competitive dynamic. The question for leaders isn't whether to adopt AI. It's whether they're building the kind of learning culture where their people know how to use it well, adapt when it changes, and grow into whatever comes next.

Effective AI upskilling programs don't just hand people tools. They create the conditions for real capability building: peer learning, personalized feedback, applied projects, and the psychological safety to experiment and fail forward.

See how Disco powers AI upskilling programs

Disco is an AI native learning platform built for human transformation. We help modern learning organizations spin up an academy and create transformative programs from their existing IP, making learning social, scalable, and personalized.

Book a demo to see how it works.