Causum announced Mars®, a decision-time AI governance platform, and launched its Key Decision-Making Authority (KDMA) program — offering the first 25 qualifying organizations a structured inventory of the decisions their AI systems make or support, at no cost.

New regulations are converging on a single requirement that most enterprises cannot yet meet: prove what decisions AI makes, what governs them, and that human oversight is operationalized — per decision, on demand. California's CCPA automated decision-making technology (ADMT) rules take effect January 1, 2026, with core compliance obligations required by January 1, 2027. The EU AI Act mandates continuous, systematic risk management for high-risk AI systems by August 2026. Legal analysis from Baker Botts concluded that under the finalized CCPA regulations, "written policies alone are now insufficient" — organizations must prove safeguards that "can withstand independent audit."

Current AI governance tools monitor models and their behavior. None formalize the decisions themselves — what each decision requires, what criteria govern it, what evidence supports it, and who holds authority over it.

"The governance market built an entire industry around watching AI after it decides. That is an autopsy, not governance," said Reza Fatahi, Ph.D., founder and CEO of Causum.

Mars enforces organizational knowledge against AI decisions before execution, using ontologically structured knowledge graphs in a specialized runtime. In benchmarking against ValuePrism — the only publicly available proxy for key decision-maker evaluation — Mars demonstrated 82% mean accuracy across 900 evaluations, outperforming hardened retrieval-augmented generation by 2.2× and zero-shot language models by 4.2×. Only 11% of organizations have agentic AI in production, according to Deloitte's 2025 Emerging Technology Trends study — a gap Causum attributes to the absence of decision-level governance infrastructure.

KDMA Program Details
The program is limited to the first 25 qualifying organizations in regulated industries. Causum creates their KDMA benchmark dataset and conducts a side-by-side comparison against existing governance tools to measure decision coverage gaps and accuracy. Participants retain full ownership of the dataset for internal use with no further obligations.