The Company Builder

Every Business Builds Three Things

Every business builds three things: the product, the company, and the business. Building company—the systems and culture that hold everything together—is the core flywheel.

Most organizations fragment this work. HR thinks about people, IT thinks about systems, Real Estate thinks about space. Employees experience these silos and handoffs as friction. 

With rigid silos between these functions, coordination becomes the work as each optimizes for its own incentives. With unified leadership, the work becomes serving employees.

This integration creates compound advantages for AI adoption. At Roblox, security decisions shape employee experience as much as collaboration tools or physical workspace. Having InfoSec at the table prevents security theater from undermining productivity.

The Trait, Not the Training

Engineering thinking can be decomposed into two things: problem-solving and curiosity. These are traits anyone can cultivate, not credentials requiring specific degrees.

When facing long-term technological shifts like AI, leaders may vacillate between optimism and pessimism. The constant? Never taking your foot off the pedal of being curious.

That mindset shapes where insight comes from. Executive forums often recycle conventional wisdom. University students working on ML and junior hires who are AI-native bring unfiltered experimentation and fresh perspectives on adoption possibilities.

A Little Discomfort Is Necessary

Most change management approaches optimize for comfort: calming communications, extensive training to prevent confusion. But a different view deserves consideration.

A little bit of discomfort is necessary. Organizations that over-engineer comfort during technological transitions slow their own progress. Letting people sit with uncertainty builds the resilience required for continuous adaptation. Anxiety signals engagement, not resistance.

Many transformations fail because they're trying to boil the ocean. Too many moving parts create friction. Better to pick one variable and hold the others constant, like deploying a new AI tool within existing processes and current team structures.

Scrappy pilots reveal real challenges faster than long-winded planning exercises. At Roblox, the empathy of people leaders combined with technical experts' efficiency enables rapid iteration based on what actually works.

Creative Games

Given the headlines about the use of AI in candidate screening, how should modern People functions think about boundaries?

Some organizations use AI extensively in recruiting while drawing clear lines. The critical distinction: avoid AI for final hiring decisions. Human judgment remains essential for choices that significantly impact people's lives and careers. AI should dramatically improve candidate experience and assessment quality leading up to those decisions.

Job requirements and descriptions remain notoriously terrible. AI agents could reduce both the burden and improve the quality of that foundational work. At Roblox, AI has helped create "one of the coolest early career experiences." Candidates play games that test characteristics such as creativity and problem-solving.

Beyond being engaging, interactive assessments resist the manipulation that plagues traditional interview tests.

The vision: AI counterintuitively enables more human bidirectional assessments between hiring managers and candidates. But integrating AI in people processes requires hard work in the underbelly of those workflows. 

As Hearst discovered when training 7,000 employees on AI, moving fast requires acknowledging employee anxiety and building local champions.

Redefining the Entry Level

Entry-level jobs have been automated away for the past 20 years. What changed is the speed. Rather than protecting entry-level roles from automation, organizations should elevate entry-level expectations with automation.

Traditional first- and second-year work can now be tool-assisted. Early-career employees should advance much faster when equipped with a learning mindset. Organizations that embrace AI-native graduates gain fresh workflows and higher output.

The leadership mindset shift? Help people do what they're passionate about by using tools to handle routine work and provide judgment-based mentorship for complex challenges.

Here's a practical filter for identifying learning capacity: Don't ask what someone is good at or bad at; ask what they're getting better at. That reveals their rate of progress, which matters far more than where they start.

‍What Can You Do on Monday?

Three actions to move toward integrated adoption:

Launch a cross-functional pilot. Identify one employee pain point that requires HR, IT, and Real Estate to solve together. Start small, start scrappy. Choose something where AI clearly helps and success is measurable within 30 days.

Schedule listening time with junior employees. Ask recent hires how they already use AI tools in their daily work. Their workflows reveal adoption opportunities that are invisible to executives. They're your leading indicators of what becomes normal.

Audit one change initiative for variable reduction. Review a current transformation effort. Changing people, process, and technology simultaneously creates confusion. Pick one variable to change while holding the others constant.

The traditional friction points—handoffs between departments, competing priorities, territorial concerns—diminish when unified leadership eliminates the coordination overhead.

These lessons apply regardless of your org chart: curiosity matters more than credentials, discomfort signals engagement, and junior employees often see what executives miss.

The question for every leader: What becomes possible when you eliminate the handoffs?