---
title: Kosuke Shirako on Why Companies Using the Same AI Reach the Same Conclusions
description: Tokyo studio SHIRO & Co. launched the Kosuke Protocol, a five-stage method for reading weak signals and second-order effects that shared AI cannot produce.
author: Darie Nani (Editor-in-Chief)
date: 2026-09-02T10:22:57.823Z
updated: 2026-09-02T10:26:24.761Z
canonical: https://www.sovereignmagazine.com/article/kosuke-protocol-shiro-co-ai-observation
image: https://cdn.nanimediahouse.com/shiro-co-kosuke-shirako-hero-v2.webp
categories: Artificial Intelligence, Business
content_type: Spotlight
region: Tokyo
publication: Sovereign Magazine
about:
  - type: Organization
    name: SHIRO & Co.
    description: "SHIRO & Co. is a Tokyo-based observation and foresight studio founded in 2025 by Kosuke Shirako. It publishes a network of public observatory sites and a bilingual journal on weak signals across society, work, bodies, technology and markets. The studio developed the Kosuke Protocol, a method for reading weak signals and second-order effects through five stages: Fragment, Observation, Connection, Meaning and Form."
    url: https://shiroand.io
    industry: Research and foresight
---

An organisation that feeds its questions to the same AI models as its competitors, trained on much the same data, should expect to [receive much the same answers](https://www.sovereignmagazine.com/article/ai-answers-too-quickly-and-it-might-be-changing-how-we-think). Kosuke Shirako has built a company on that problem. On 25 August his Tokyo studio, SHIRO & Co., introduced Kosuke Protocol, a method for turning small observed changes in society, markets and human behaviour into products, enterprise systems and business decisions, and it is now opening discussions with enterprises, investors and partners on private implementations, licensing and joint development.

SHIRO & Co. describes itself as an independent Observation & Protocol Studio. Rather than shipping unrelated AI applications, it runs an observatory network across markets, manufacturing, education, labour, climate, risk, culture and everyday life, and connects what those observatories record. The protocol moves through five stages: fragment, observation, connection, meaning and form. The final stage is deliberately open: the same observed material might become a market-intelligence interface, a proposal system, a consequence simulator, an educational environment or a piece of narrative writing.

> "I believe the next source of competitive advantage will not simply be better AI. It will be a distinctive way of observing and interpreting change. That is why I wanted to build a protocol rather than a single product: the product may change, but an organization's way of reading change should become a lasting capability."
> — Kosuke Shirako, Founder, SHIRO & Co.

Research over the past five years supports the premise. Kleinberg and Raghavan's 2021 PNAS paper on algorithmic monoculture showed that when many actors adopt the same algorithm, collective outcomes can worsen even when the algorithm is individually the best available. A 2022 NeurIPS study found that shared base models homogenise outcomes across firms that deploy them. In 2024, [research in Science Advances](https://www.science.org/doi/10.1126/sciadv.adn5290) found that generative AI raised the quality of individual writers' work while narrowing the collective range of what was produced, and Jay Barney and Martin Reeves argued in [Harvard Business Review](https://hbr.org/2024/09/ai-wont-give-you-a-new-sustainable-advantage) that AI adopted by everyone [cannot by itself be a source of advantage](https://www.sovereignmagazine.com/article/what-are-the-next-steps-for-investors-and-execs-as-nobel-winners-say-tech-growth-needs-compet) for anyone.

## AI Model Comparison as Market Intelligence

One of the studio's enterprise implementations, Machine-Constructed Markets, applies the point directly. It compares how different AI systems associate a company with known, adjacent and emerging markets, setting the readings of OpenAI, Claude and Gemini side by side to show where the machine intelligences agree, where they diverge and where one of them has identified a category the others have not. The company is explicit that these observations are not statements of fact. They are inputs for corporate planning, positioning, partnership discovery and M&A hypothesis development: a picture of how the machines that increasingly mediate markets currently construct yours.

A second implementation, Signal to Proposal, converts a detected market change into a company-specific brief through five stages of its own, from signal through structure and meaning to opportunity and proposal. The studio pitches it at manufacturing, industrial technology, B2B software, certification, risk and finance.

## Usable Hours: Extreme Heat and the Working Day in Kochi

The clearest demonstration of the method comes from the studio's Climate Observatory, and it starts with a rule now in force in one Japanese prefecture. Since March 2025, Kochi's Cool Work Time guideline for public construction works has split the summer working day into 8am to 11am and 2pm to 5pm, suspending work through the hottest three hours. Because that window includes the conventional midday break, the effective loss is roughly two working hours: a conventional eight-hour outdoor day becomes six. The International Labour Organization projects that heat stress will remove 2.2 per cent of global working hours by 2030, the equivalent of 80 million full-time jobs, and since June 2025 every Japanese employer has been under a mandatory heatstroke-countermeasure ordinance with criminal penalties attached.

![SHIRO & Co. Climate Temporal Compression interface: a standard workday and a climate-adapted day plotted hour by hour, with heat-constrained hours highlighted and usable hours lost](https://cdn.nanimediahouse.com/shiro-co-climate-temporal-compression-full-v2.webp)
*Photo: SHIRO & Co.*

"In one case we examined in Kochi, Japan, extreme-heat measures effectively remove roughly two hours from a conventional eight-hour working day — changing the question from 'How hot is it?' to 'What happens to productivity, project schedules, labor systems, and regional competitiveness when usable time contracts?'" Shirako said.

The studio calls the underlying variable usable human time, and treats it as a business quantity: something that contracts, moves project schedules, changes automation demand and redraws regional competitiveness. Its own analysis of environment-ministry heat data this summer found Kochi losing a median hour of usable outdoor time per day across July and August, with single days losing six. SHIRO & Co. presents those figures as experimental observation rather than an occupational-safety determination.

## Second-Order Effects and Where the Costs of a Decision Move

A third framework, Entangled Society, works on the consequences side of the same discipline. Most interventions that succeed somewhere create second-order effects somewhere else: AI adoption, a supply-chain redesign or a labour reform produces benefits in one place while cost, dependency or risk move to another. The framework's consequence simulator asks a fixed set of questions of any intervention. Who benefits? Who carries the burden? Who decides? Who cannot exit? Where does the cost move? The aim is not to predict a single future or to grade a decision as good or bad, but to surface the unintended consequences before commitments harden into things that are difficult to reverse.

## An Observatory Network Run in Public

What separates the studio from the strategic foresight and horizon scanning trade is that its raw material is published. A growing network of observatory sites is live and readable by anyone, covering market signals, the education-to-employment gap, scam folklore in the AI era, intimacy and technology, the culture of AI-generated things and more, alongside a bilingual journal of observations in English and Japanese. The network functions as a working knowledge management system that anyone can read: consultancies in the same territory sell their horizon scanning as decks and dashboards, while SHIRO & Co. demonstrates its weak-signal work in the open and sells the deeper layer.

That deeper layer is Private Protocol, the studio's commercial offer: implementations that combine the protocol with a company's own knowledge, decision criteria, customer context and observational data, so that the machinery for turning weak signals into strategy belongs to the company rather than to a shared model. Discussions on offer run from proof-of-concept programmes and licensing to co-development and capital alliances.

## From a Cross-Industry Career to a Named Method

Shirako says the protocol distils more than twenty years leading marketing and business development across ten or more companies, spanning telecommunications, broadcasting, semiconductors, industrial software, IoT, SaaS, financial data, and certification and risk services. Moving repeatedly between technology, market, customer and organisation, he says, built the cross-layer reading of change that the protocol now formalises.

"The enduring asset is not my résumé — it is the system that allows observations to travel across domains and accumulate into meaning. That is difficult to replicate simply by adding another AI model or research team," he said.

The studio has filed a trademark application for Kosuke Protocol in Japan and says it is global by design, with most of its interfaces published in English because the partnerships it is pursuing are international rather than domestic. Detailed observatory interfaces and prototypes are disclosed selectively, some through paid engagement.

## FAQ

**Q: What does the term "weak signals" mean in business strategy?**
Weak signals are early, easily missed indicators of change: a new regulation, a shift in language, an emerging technology or an overlooked behaviour. Individually they look insignificant. Kosuke Protocol is built on the premise that fragments observed over time and connected across domains can reveal a new market, decision system or human experience before it has a name.

**Q: What is a second-order effect?**
A second-order effect is the consequence of a consequence: the change an intervention causes beyond its immediate, intended result. SHIRO & Co.'s Entangled Society framework maps these by asking who benefits from a decision, who carries the burden, who decides, who cannot exit and where the cost moves, before commitments become difficult to reverse.

**Q: What is an example of a strategic foresight method?**
Kosuke Protocol is one: a five-stage method that moves from fragment through observation, connection and meaning to form. Unlike scenario planning exercises that end in a report, its final stage is deliberately open, so the same observed material can become a market analysis, a business proposal, a simulation or a product.

**About SHIRO & Co.**

SHIRO & Co. is a Tokyo-based observation and foresight studio founded in 2025 by Kosuke Shirako. It publishes a network of public observatory sites and a bilingual journal on weak signals across society, work, bodies, technology and markets. The studio developed the Kosuke Protocol, a method for reading weak signals and second-order effects through five stages: Fragment, Observation, Connection, Meaning and Form.

[Website](https://shiroand.io)