The Intellectual Foundation
of QuantFrame.
QuantFrame is built on a decade of research at the intersection of institutional economics, quantitative finance, and AI governance. The core insight: authority migrates into automated systems faster than governance mechanisms can maintain intervention rights.
Human decision-making operates within cognitive limits — not because we are irrational, but because the world is too complex for perfect optimization. QuantFrame extends this insight to autonomous systems: AI agents are also bounded, but the limits are architectural, not cognitive.
Institutions shape the rules within which decisions are made. When those rules are unclear, incomplete, or absent, actors default to informal norms — and accountability breaks down. QuantFrame treats the IPS as the institutional rule set for AI action within a financial firm.
Portfolio theory gave us the formal language of risk and return trade-offs. The QuantFrame Frontier extends this framework by introducing governance quality as a third state variable — showing that the achievable frontier is not fixed, but shifts with the quality of the governance regime.
When an organization deploys an autonomous system, it delegates agency to that system. The system acts — executes trades, generates recommendations, sizes positions — on behalf of the organization and its clients. The question is not whether the system is capable. It is whether the governance architecture can maintain meaningful intervention rights when it matters.
Bounded Agency is the condition in which autonomous action is constrained by a governance architecture that preserves human oversight, reclaimability, and accountability. The failure mode is not malicious AI — it is the gradual migration of authority into systems that operate faster than human intervention can track.
"The problem is not AI capability. It is the gap between the speed at which AI systems make decisions and the speed at which humans can meaningfully intervene."
A firm's governance architecture is either structurally sound enough to support AI deployment — or it is not. QuantFrame makes that determination measurable.
The BAL Score measures an organization's structural capacity to govern autonomous AI systems. It is a process score — not a returns score. The same IPS and the same model portfolio can produce different BAL Scores across different firms, because execution quality and governance discipline vary.
Pilot access. The full BAL Score methodology, primitive definitions, and scoring architecture are disclosed to pilot firms under NDA. Request access →
Markowitz gave us the efficient frontier — the set of optimal portfolios maximizing return for a given level of risk. The QuantFrame Frontier introduces a third variable that classical portfolio theory does not account for.
The result is a family of governance-adjusted efficient frontiers. Each BAL Score regime produces a distinct achievable frontier. The full framework — including the formal construction, the governance state variable, and its empirical implications — is the subject of ongoing research and is disclosed to pilot firms.
"Governance does not generate alpha. It eliminates execution drag — and execution drag, compounded over time, is the silent destroyer of client outcomes."
Research. The foundational Bounded Agency paper is available on SSRN. Read the paper →
Your IPS is not just a compliance document. It is the governance primitive from which every AI decision in your firm should derive its authority. QuantFrame operationalizes it.
Strategy fidelity — whether your fund does what your strategy document says — is now measurable at the portfolio level, across every AI action, in real time.
The BAL Score introduces a governance vocabulary for AI in finance that regulation is still developing. It gives examiners a structured lens for assessing governance quality, not just intent.
Accountability requires auditability. The QuantFrame framework gives boards a governance artifact — not a verbal assurance — that autonomous systems operated within mandate.
Prof. P.K. Prasanna Kumar
QuantFrame is invitation-only at launch. Request pilot access and generate your first portfolio BAL Score.