1

A Managed Heart? Navigating Emotions Under Algorithmic Management

by Hadi Shaheen, assistant professor of management

“I’ve learned that people will forget what you said, people will forget what you did, but people will never forget how you made them feel.” — Maya Angelou.

That quote sits at the center of this teaching exercise, which is part of the recent AI projects and grant work at the School of Business, but rather than people, I used AI to simulate managers and provoke the emotions that shape workplace decisions.

I was inspired by an informal discussion on AI and pedagogy with Dean Paul and my colleague Maryam Mahdikhani during our outing following the SOB meeting at BarTaco.

I created and introduced students to two AI “bosses” with contrasting styles and asked them to pitch a venture idea to each. Boss A embodied the style of Steve Jobs or Marissa Mayer at their peak: autocratic, perfectionist, brutally candid. Boss A also echoed Andy Grove’s paranoia about execution and Jeff Bezos’s bar-raising, metric-obsessed critique. Boss B, on the other hand, channeled Indra Nooyi’s empathetic, growth-mindset leadership, a warm, transformational, challenge-with-care attitude. It blended Reed Hastings’s freedom-and-responsibility experimentation with Ed Catmull’s psychological safety.

I asked students to notice as they pitched not just what they thought, but how they felt and what that did to their choices. I also asked them to decipher each boss’s leadership style and figure out ways to navigate it to get what they want.

The Pedagogy

AI can be more than a shortcut for efficiency; it can be a simulator. Students can experience what it’s like to work under different team managers. It helps students do two things they’ll need in real workplaces: (1) read the boss: interpret a manager’s cues, cadence, and thresholds and (2) regulate themselves: name their own emotional state and choose a productive action rather than simply reacting.

Practically, this matters because algorithmic management isn’t hypothetical. It’s already here. In fact, it was here for a while. Rideshare and delivery platforms route workers, assign tasks, rate performance, and can fire workers based on metrics. Navigation apps nudge behaviour in real time. On a personal note, my own Garmin watch often tells me I’m “unproductive” in my training routines and adjusts my running targets. I follow its orders, but it sometimes makes me feel helpless. Whether we like it or not, algorithms shape our feelings and, therefore, our decisions. AI in the classroom can create a safer place to learn how to notice those pressures and work with them.

The Setup

Students pitched, received rapid feedback, revised, and re-pitched in short cycles to the two custom GPTs. After each exchange, they noted a quick “felt snapshot” (e.g., energized, tense, confident, defensive) and a behaviour choice (e.g. shrink scope, reframe, test, escalate).

How It Unfolded in the Room

Patterns emerged quickly across students:

Boss A- Students reported sharper thinking: naming “adoption friction, integration risks, execution complexity.” One wrote: “Ruthless. Cuts through the BS. Makes you think fast.” Another: “It’s not stuff I want to hear but definitely need to think about.” The costs were real: “Makes me feel stupid,” “rage-baiting,” and “no answer is good enough.” Yet some students discovered that pushing back productively against Boss A could catalyze sharper pivots and better business models: “We argued, then pivoted to a better, narrower business.”

Boss B- Comments included “very positive and constructive” and “reinforced my idea and showed how to elevate it.” Several appreciated how Boss B organized thoughts into clearer categories. But students also flagged over-scaffolding: “It does the work for you,” “good to get something done, but detracts from creativity,” and “helpful maybe, but unrealistic if unchallenged.”

Overall preferences split by person and phase. Some wanted Boss A’s pressure to “grow your brain through critical thinking and fast problem-solving.” Others preferred Boss B’s encouragement: “I don’t benefit from negative reinforcement, so A made me want to do the work less.”

Students’ discussions also triggered an interesting strategy hypothesis: workplace confrontation narrows exploration but raises agency; workplace conformity expands exploration but reduces agency.

What Students Practiced

• Reading the room: spotting a leader’s style, thresholds, and “feeling rules,” then tailoring the pitch.
• Navigating workplace emotion: noticing one’s state and choosing an intentional next move.
• Ambidexterity in practice: students learn to diverge ideas under Boss B and converge under Boss A, switching modes on demand.”.

Next semester, I’ll expand the roster of boss archetypes, starting with the Data-First Skeptic Boss! The aim isn’t to crown a favourite boss, but to help students read the manager and self-regulate so the same idea lands under different, real leadership styles. I’ll share the lineup as I develop it more, and I’d welcome collaborators, especially colleagues in psychology and organizational behaviour.

Image created with Google Gemini 2.5 Flash