For L&D Evaluators
The Research Behind LeaderCoreAI
The research behind LeaderCoreAI is the first thing serious evaluators ask about. Is it actually grounded in leadership research, or is it a general-purpose chatbot with a leadership label stuck on top? It's a fair question. The market is full of tools that added a flimsy general-purpose system prompt and then called it AI roleplay, or AI leadership training. LeaderCoreAI goes much further than that. This page explains what our scoring is actually built on.
Short Answer
LeaderCoreAI's scoring is grounded in established leadership research. What sets it apart in this category: the grounding is competency-specific. Each simulation is scored against the literature that defines its specific competency, and matches only observed behaviour against those competencies.
The Shared Foundation
Every LeaderCore product rests on the same principle: scoring is grounded in established, peer-tested research for the specific competency being assessed, not one generic leadership framework stretched across every product. Which research applies depends on the product, because the competency does too. This principle is the price of entry - any serious leadership development tool should be able to say the same.
For example, the leadership-development products - LeaderCore: Inspire (including its VP-level scenario variant) and LeaderCore: New Manager - are grounded in three bodies of research:
The basis of the Leadership Practices Inventory (LPI), one of the most widely used and validated 360-degree leadership assessments. It frames leadership as observable, learnable behavior organized around five practices: Model the Way, Inspire a Shared Vision, Challenge the Process, Enable Others to Act, and Encourage the Heart.
Goleman's framework for self-awareness and self-regulation under pressure. It shapes how simulations read composure and interpersonal judgment in high-stakes conversations: not just what a leader decided, but how they handled the moment.
Edmondson's research on the conditions under which teams speak up, admit mistakes, and learn instead of staying quiet. It shapes how simulations evaluate whether a leader's behavior invites honesty from the people around them or shuts it down.
Where It Gets Specific
LeaderCore: Influence makes the principle unmistakable, because it draws on a completely different body of research than the leadership products above - not a variation on Kouzes & Posner, Goleman, or Edmondson, but the literature that actually defines influence as its own competency.
LeaderCore: Influence
Influence isn't scored against the same rubric as a feedback conversation or a crisis scenario. It draws on:
- Robert Cialdini's principles of influence and persuasion - reciprocity, commitment and consistency, social proof, authority, liking, and scarcity. The mechanisms that make persuasion work or fail.
- Cohen & Bradford's Influence Without Authority model - the exchange of “currencies” (resources, information, position, relationship, personal standing) that let someone build cooperation without formal power over the other person.
When a manager runs a LeaderCore: Influence simulation, their score reflects whether they applied these specific mechanisms, not whether they hit a generic “communicated clearly” checkbox.
That's the line between a purpose-built assessment and a chatbot with a leadership label. A general-purpose model can hold a conversation about influence. It doesn't know that influence, as a competency, has its own defining literature, and that scoring it well means scoring against that literature rather than against leadership in general. The research that defines a competency is the research we score it against.
Not Decorative: How the Research Gets Enforced
Citing the right researchers only means something if it shows up in how a session actually gets graded. Two mechanics enforce it:
An invariant scoring rubric
The competency framework for a given simulation is the standard, the same standard, regardless of industry, seniority, or scenario difficulty. LeaderCore: Inspire applies the same Kouzes & Posner, Goleman, and Edmondson-based criteria whether the scenario is set at the middle-manager level or in one of its VP-level scenarios. LeaderCore: Influence applies the same Cialdini- and Cohen & Bradford-based criteria no matter which industry the scenario is set in. Neither rubric loosens because the conversation got harder.
Evidence-cited scoring
Every score traces back to specific things the learner actually said or did in the simulation, not a holistic impression. If a manager is marked down on reciprocity, the feedback points to the moment in the transcript where they missed it.
Combined with the competency-specific grounding above, that's the full claim: the frameworks that define the competency, applied at a consistent standard, evidenced against what the learner actually did.
In Summary
The Leadership Practices Inventory, Goleman's and Edmondson's research, Cialdini's work, and Cohen & Bradford's model are established bodies of research in their own right. LeaderCoreAI is grounded in them and applies them inside a simulation environment. The scoring logic behind each simulation is built on the literature that defines the competency it tests, applied consistently, and evidenced against real behavior in the session.