QommonsAI Equipped with Quantitative Visualization of Business Reduction Effects, Input/Output Log Management, and Super Administrator Privileges — "Not Just an Introduction." Towards an Era of "Proving the Effects of Generative AI with Numbers"

Polimill Inc. (Headquarters: Minato-ku, Tokyo, Representative Directors: Ayame Ito / Nonoka Taniguchi), a top runner in administrative AX, will begin offering "Effect Measurement Dashboard (Quantitative Calculation of Business Reduction Effects)," "User Log Management (Recording, Storage, and CSV Output of Input/Output Logs)," and "Super Administrator Privileges (Permission Control with 3-Tier Roles)" in an update scheduled for April 1, 2026, for its generative AI for government, "QommonsAI (Commons AI)."

■ "Just saying 'we are using it' is no longer enough" ─ The "Quantitative Accountability" Required for Generative AI Utilization

As generative AI adoption progresses in municipalities nationwide, a new challenge is emerging: "How to explain the effects of adoption."

In discussions in assemblies, budget requests for the following fiscal year, and reports to information security committees, qualitative answers like "we are using generative AI" are becoming insufficient. What is required is a system that can show in numbers "who, in which department, how they used it, and how much business time was reduced."

At the same time, without a system for managing input/output content and storing logs, the effectiveness of operational rules cannot be verified. Furthermore, if it's not clear "who can access" this highly confidential log information, governance itself cannot be established.

To solidify the "continuable form" of generative AI utilization, the three wheels of "visualization of effects," "transparency of operation," and "control of privileges" are indispensable.

■ 3-Tier User Roles ─ Controlling "Who Can Do What" with Organizational Rules

With this update, QommonsAI will introduce "Super Administrator / Administrator / Member" - a 3-tier user role system.

In the operation of generative AI, access logs, input/output content, and access rights to internal knowledge bases are extremely confidential information. "Who can view the input content of all employees?" "Who controls the scope of sharing internal documents?" Continuing operations without clear answers to these questions can lead to significant governance risks.

QommonsAI addresses this challenge with an approach of "hierarchical authorization."

Viewing and downloading input/output logs is limited to Super Administrators only; the "User Logs" menu will not be displayed for users other than Super Administrators.

By technically controlling access to both logs (the "most sensitive information") and the knowledge base (the "most extensive information"), governance is realized where "only those who should see can see, and only those who should manage can manage."

This applies the "principle of least privilege" from the internal information security policy to AI operations, providing clear grounds for explanation in reports to information security committees and audits.

■ Effect Measurement Dashboard ─ An Administrator Screen to "Talk About Generative AI's Introduction Effects with Numbers"

The QommonsAI Effect Measurement Dashboard is not just a usage statistics screen. It quantitatively visualizes the business reduction effects brought about by generative AI from four perspectives: "How much is it being used?", "Which functions are effective?", "How much reduction has been achieved in which tasks?", and "Where has it been adopted, and where are the challenges?".

Indicators necessary to "show the introduction effects with numbers," such as utilization status by department and estimated reduction time by business category, are consolidated on the administrator screen. It also allows for grasping usage disparities and underutilized areas, directly informing the prioritization of additional training and internal deployment measures.

Furthermore, it supports automatic generation of monthly reports. This reduces the reporting workload for administrators and can be used directly as quantitative evidence in assembly reports and budget requests.

■ User Log Management ─ Storing All Input/Output Records in Domestic Regions

The User Log function comprehensively accumulates and manages logs of "who, when, what was input, and to what extent it was used." Administrators can view logs on the dashboard and download them in CSV format.

User data and logs are all stored in domestic regions, ensuring peace of mind from a data sovereignty perspective. Logs are retained for the entire contract period and are completely deleted after an agreed-upon period following the end of the usage period.

■ From "Proof of Introduction Effects" to "PDCA for Internal Deployment" ─ Assumed Use Cases

Information Policy Division: Utilize monthly reports from the Effect Measurement Dashboard to present quantitative data such as "Estimated reduction time XX hours/month" in assembly responses and budget request materials.

General Affairs Division / Administrative Reform Promotion Office: Grasp departmental utilization rates and underutilized departments to determine priorities for additional training and internal deployment plans.

Human Resources Division: Track changes in utilization rates over time after training implementation and use for quantitative evaluation of training effectiveness.

Audit Commissioner's Secretariat: Utilize CSV output of input/output logs as verification material for operational status in information security audits.

Information Security Administrator: Centrally manage logs with Super Administrator privileges, grasp the entire history of "who entered what and when." Clearly indicate the status of access control in reports to the security committee.

All Staff: Support the shift from "we are using it" to "this much effect is being achieved," accelerating internal consensus-building for generative AI utilization.

■ QommonsAI Aims for "Explainable AI Operations"

QommonsAI continues to implement functions that provide a foundation for municipalities to "use generative AI with peace of mind." Following the "3-layer defense" of "personal information automatic detection, forbidden word blocking, and detection logs" announced on March 28, the current "Effect Measurement Dashboard," "User Log Management," and "Super Administrator Privileges" support the sustainable promotion of generative AI utilization in government through the integration of security, governance, and effect demonstration.

◾️About QommonsAI

QommonsAI, a generative AI for government developed, provided, and operated by Polimill Inc., supports the resolution of municipal issues on an evidence basis, drawing on data from tens of millions of domestic and international laws, policies, papers, and municipal case studies. As of March 2026, it is used by over 700 municipalities nationwide and approximately 250,000 people. It is utilized in a wide range of tasks, including assembly response, policy planning, resident services, and public relations. It is equipped with 12 models carefully selected from the four major companies OpenAI, Anthropic, Google, and Preferred Networks, providing an environment where users can select the optimal AI for their needs. It is offered directly to all municipalities without intermediaries, and since the service release, free on-site introductory and basic/intermediate training has been provided to all adopting municipalities. Furthermore, it operates with an agile development system that reflects requests from municipal sites into the service in as little as a few days, continuously delivering "functions needed now" for government at overwhelming speed.

Official Website: https://info.qommons.ai/

◾️Company Overview

Company Name: Polimill Inc.

Location: Minato-ku, Tokyo

Representatives: Representative Directors Ayame Ito / Nonoka Taniguchi

Business Description: Planning, development, and operation of generative AI for government "QommonsAI," digital democracy platform "Surfvote," etc.

Corporate Site: https://polimill.jp/