Every musician who walks into a meeting with a concert promoter has spent thousands of hours preparing. Not for the meeting — for the stage. They've rehearsed scales, études, sonatas, concertos. They've performed in masterclasses, competitions, and recitals. They've been trained, evaluated, and refined by some of the best teachers in the world.

But for the conversation that determines whether they actually get to perform? Zero preparation. Zero practice. Zero training.

They walk in and wing it. Sound familiar?

The myth that talent speaks for itself

Classical musicians grow up with a promise: if you practice hard enough, if you're good enough, the career will follow. An orchestra will hire you. A promoter will book you. An agent will find you.

For a tiny percentage, that's true. For the vast majority — including many extraordinarily talented musicians — it isn't. The music industry, like every other industry, runs on conversations. Conversations about programming, about fees, about artistic vision, about logistics. Conversations where a musician needs to understand what a promoter needs before presenting what they have to offer.

Most musicians never learn this. They treat meetings with promoters like auditions — performing a rehearsed pitch while the promoter passively judges. The result is predictable: the musician talks about themselves, the promoter listens politely, and nothing happens.

This is the rehearsal gap — applied to an entirely different profession.

Building the scenario

When the University of Music Karlsruhe (HfM Karlsruhe) asked us to develop a training scenario for their students, we didn't start with the technology. We started with the problem.

Musicians in Germany face a brutal economic reality. The average freelance musician earns €2,660 net per month. One in five earns less than €1,500. The new BKM minimum fee — €350 per day, effective March 2026 — is a floor that many promoters still try to undercut. Most musicians have never been taught to negotiate, to discuss money with clarity, or to position themselves as professional partners rather than grateful supplicants.

The issue isn't confidence in the abstract. It's that musicians have never practiced the conversation. They've never experienced what happens when a promoter offers "exposure" instead of a fee and had to respond in real time. They've never navigated the moment when a prestigious festival director gives them exactly 15 minutes and shows no patience for a generic pitch.

So we built a scenario with four learning objectives, weighted by importance:

Discovery Before You Pitch (35%) — ask genuine questions about the promoter's situation before presenting your project. Most musicians skip this entirely.

Connect Your Project to Their Needs (25%) — frame your artistic concept as a response to what you learned, not a generic pitch deck delivered regardless of context.

Navigate Fee and Logistics as Partners (20%) — discuss money with clarity. Know your BATNA (best alternative to a negotiated agreement). Understand the promoter's business model. Don't leave money on the table — and don't beg.

Position as Collaborator, Not Supplicant (20%) — maintain professional self-respect throughout. Adult-to-Adult, not Parent-to-Child. This is where Transactional Analysis meets artistic partnership.

Six personas, six realities

We designed six AI promoter personas, each representing a different professional situation musicians actually encounter — graduated from Easy to Very Challenging.

There's Paco, the luxury resort manager on Lake Como who offers a beautiful venue and wealthy audience but no cash fee. He tests whether a musician can hold their value when offered "the experience" instead of payment. Students who advocate for themselves warmly and firmly unlock a travel contribution. Students who accept the first offer get nothing.

There's Elena, the burned-out Nordic festival programmer who's distracted, checking her phone, giving one-word answers. Most students get a polite brush-off in eight minutes. Students who ask about her situation — her challenges, her budget fights, her vision — discover a passionate curator hungry for genuine artistic conversation. The meeting extends to twenty minutes or more.

And there's Julia, the director of a prestigious London agency who uses brevity and status as power signals. She tests whether a musician can maintain professional composure under sustained pressure. Students who hold their ground with grace may learn she's actually looking for the right artist for her roster.

Every persona uses a Trust Score mechanism. Good technique — asking before telling, genuine curiosity, building on what the promoter shares — increases trust and unlocks richer responses, hidden information, and better outcomes. Poor technique — pitching without listening, supplicant language, giving advice without asking — produces visibly shorter, flatter, more disengaged responses.

The difference between a good interaction and a poor one isn't told to the student. It's felt — in the texture of what they get back. Just like in a real meeting.

What 61 sessions taught us

We've now run 61 sessions with real students. The data tells a clear story.

The average score across all sessions is 63 out of 100. Scores range from 24 to 95. The distribution: 6% Excellent, 36% Good, 44% Needs Improvement, 14% Poor.

That distribution matters. It means the scenario provides meaningful challenge without being discouraging. Students aren't acing it on the first try — which would mean the scenario is too easy to produce learning. And they're not failing catastrophically — which would mean it's too hard to be useful.

The most frequently chosen persona is Maia, the Berlin festival director who only responds to genuine creative collaboration (20 sessions). Second is Paco, the Lake Como resort manager (16 sessions). Third — and this is telling — is Julia, the most challenging persona (12 sessions). Students are drawn to the difficult conversations, not the easy ones. They want to test themselves against real pressure.

What this has to do with leadership

Everything.

Replace "musician" with "director" and "promoter" with "board member," and the scenario is identical. A leader walks into a strategic conversation having prepared the content — the numbers, the slides, the talking points — but having never practiced the conversation. The moment someone pushes back, changes the subject, goes silent, or doesn't follow the script, the leader is on their own.

The four learning objectives we designed for musicians apply directly to any leadership conversation: discover before you pitch, connect your proposal to their needs, navigate negotiation as partners, and position as a collaborator rather than someone asking for permission.

The persona design works the same way too. In our corporate scenarios, we build personas that reflect the specific dynamics of each organization — the resistant team member, the skeptical board director, the burned-out peer who's checked out. They use the same Trust Score architecture: good conversational technique unlocks depth and progress; poor technique produces flat, disengaged responses. Just like real people.

And the anti-sycophancy design that makes the music scenario work — personas that don't cave after two sentences, that maintain realistic tension throughout — is the same architecture that Stanford research (Shaikh et al., CHI 2024) identified as essential for producing behavioral change. Standard AI is too agreeable to be useful for practice. We build AI that pushes back.

Why we're writing about this

We could keep this quiet. Classical music sales is a niche. It doesn't sound like our core market.

But it demonstrates something that a corporate case study can't: the depth of scenario design we bring to every engagement. We didn't adapt a template for Karlsruhe. We built six personas from scratch, each with hidden depths, trust mechanics, and calibrated difficulty. We grounded the scenario in specific economic data, specific negotiation frameworks, and specific learning objectives weighted by importance.

That's exactly what we do for every organization we work with. Whether the conversation is about concert fees or corporate strategy, the methodology is the same: start from learning objectives, build personas that reflect real dynamics, design for behavioral change — not just knowledge transfer.

Because the question is the same for a musician meeting a promoter, a director meeting a board, or a manager giving difficult feedback:

You know what to say. Can you say it when it counts?

If your people are facing conversations they've prepared for but never practiced, let's talk.

References

Shaikh, O., Chai, V., Gelfand, M. J., Yang, D., & Bernstein, M. S. (2024). Rehearsal: Simulating conflict to teach conflict resolution. Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI '24). ACM.

MIZ Berufsmusikstudie (2025). Einkommens- und Beschäftigungssituation freiberuflicher Musiker:innen in Deutschland.

Why We Built an AI Roleplay for Classical Musicians — And What It Taught Us About Leadership Training