We are pleased to announce the development and initial deployment of our proprietary architecture, “DDIM™ (Decision Data Integrity Model),” which comprehensively evaluates the validity, transparency, and accountability of decisions involving both humans and AI.

This technology provides an evaluation framework that enables the structural and mathematical verification of “decision quality” and “accountability”—essential considerations in today’s era of advanced DX and AI adoption.
We have filed patent applications for the core technologies within this architecture.

■ Background

As DX accelerates and AI implementation progresses, many companies and government agencies face the challenge of
the challenge of being unable to sufficiently explain “why that judgment was reached.”

・AI inference processes are black-boxed
・Project decisions rely on personal biases and tacit knowledge
・The basis for management decisions is fragmented, making consistency verification difficult
・There are no structural means to fulfill accountability

Explainable AI (XAI) contributes to explaining AI outputs, but no mechanism existed to evaluate the “integrity” of the entire decision-making process, including human involvement.

Massive Act identified this structural gap and defined a new concept: “Decision-Making Integrity.”
We designed DDIM™ as the architecture enabling its evaluation.

■ Technical Overview: DDIM™ (Decision Integrity Assessment AI)

Name: Decision Integrity Assessment AI System (Digital Decision Integrity Model)
Common Name: DDIM™

DDIM™ views decision-making not as an outcome, but as a structured entity comprising “Claims, Evidence, Counterarguments, and Sources.”
AI organizes these as a directed graph and evaluates them along the following three axes:

・SV: Validity
Evaluates the logical consistency of the judgment and the strength of its evidence.

・T: Transparency
Evaluates the traceability and disclosure level of the decision-making process.

・A: Accountability
Evaluates role allocation, responsibility assignment, and re-evaluation feasibility.

Using the integrated DDIM Score metric, it quantitatively determines:
“How trustworthy is this decision?”
Furthermore, it functions as an evaluation × improvement engine that automates the decision improvement cycle via AI, including:

・Automatic extraction of high-impact factors
・Identification of issues like insufficient grounds or bias
・Automatic generation of improvement proposals using generative AI (RAG)

■ Positioning of the Patent Application

This patent application establishes an architecture for scientifically evaluating the legitimacy of decision-making, addressing the societal demand for “accountability and transparency” in the era of DX and AI utilization.

DDIM™ is intended for use in the following domains:

・Quality assurance for management decisions
・AI auditing and model governance
・Transparency in administrative decision-making processes
・Strengthening accountability in healthcare and public sectors

This technology is being jointly researched and developed with the Japan Digital Advisory Strategy Council (JDACs), a general incorporated association that verifies and advises on DX/AI investment decisions from a third-party perspective.
Future plans include integration with third-party review systems and advisory certification programs.

■ Future Outlook

Massive Act will further strengthen its support for “Transparent DX” and “Responsible AI Utilization” centered around DDIM™.

Furthermore, we will work to build a “Circular Decision-Making Model” leveraging synergies with our customer LTV prediction AI, “MACT INTELLIGENCE™.” We will continue technological development aiming to realize a society where decision-making is more transparent, more accurate, and more trustworthy.