Human–AI collaboration
- Define boundaries between human judgment and model output.
- Establish reusable flows, prompts, and review loops.
- Create “protocols” instead of one-off clever prompts.
Fragment Practice — a small thinking studio for human–AI work.
We help teams design human–AI protocols, document structures, and work rhythmsgrounded in how people actually think, write, and decide. Our entry point is not tooling or dashboards—it’s the small, scattered fragments you already produce every day: meeting notes, partial ideas, logs, drafts, tensions, and unspoken assumptions.
From those fragments, we help teams build quiet, sustainable operating systems: workflows that reduce cognitive noise, strengthen decisions, and allow AI to contribute without eroding human judgment or shared context.
This page provides a concise English overview for practitioners, researchers, and founders. The detailed service pages and research archive are currently available in Japanese.
What we are — a studio for structure, protocols, and reasoning
Fragment Practice operates at the intersection of systems thinking, writing, and applied AI. We treat documents, conversations, and AI interactions as part of a single knowledge system—and design the structures that allow them to reinforce, rather than undermine, each other.
Document & workflow design
- Design note/decision structures that survive scale.
- Connect meeting notes → AI summaries → actions → documentation.
- Reduce drift between text, tools, teams, and intentions.
Governance & operational clarity
- Align AI usage with security and BCP practices.
- Handle sensitive data via clear rhythms and boundaries.
- Replace “firefighting” with quiet, preventative structure.
What we do — Services, Research, and Fragment
Our work spans three connected modes. Teams often start in one and end up using all three: applied consulting (Services), structural research (Research & Writing), and our text-first notebook (Fragment).
Services — hands-on collaboration
We work directly inside teams to design workflows, AI usage, and document systems that reflect how you actually operate.
- Map existing fragments and diagnose friction.
- Prototype small protocol changes without new tools.
- Co-design sustainable work rhythms and document flows.
Research & Writing — structural notes
We turn recurring patterns from projects into reusable frameworks, notes, and write-ups that teams can rely on.
- Cross-project pattern extraction & conceptual framing.
- Prism/QFS/Fragment layer documentation.
- ZINEs, technical essays, and applied research logs.
Fragment — our notebook
Fragment is a calm, AI-native notebook built on Markdown + small YAML blocks. It is the page model that supports both our consulting and research work.
- One page = one Fragment (a unit of thinking).
- Prism/Scene/Flow metadata for AI-aware reading.
- Live examples at fragment.place.
How we work — small loops, structural change
We don’t ask for polished plans. You can start with a messy document, a complicated decision, or “something feels off.” We iterate in small, evidence-based loops:
Listen & Map
We read what already exists—notes, logs, drafts, Slack threads—and map the real constraints and decision points in your system.
Design & Sketch
Without adding tooling, we prototype the smallest useful protocol change: templates, flows, YAML roles, checklists, or reading modes for AI.
Write & Embed
We convert what worked into durable notes, internal docs, or repeatable processes—building a backbone your team can return to.
Principles — the constraints we design within
These principles shape how we write, design, and collaborate. They’re intentionally simple and operational.
Quiet rhythm
Good reasoning needs slack. We design workflows that reduce cognitive noise rather than add dashboards and notifications.
Fragments → Structure
Half-formed notes often contain the real signal. We treat fragments as inputs to structure, not failures to finish writing.
Clear boundaries
Human vs AI responsibility must be explicit. Ethics and governance emerge from boundary-setting, not after-the-fact controls.
Tools as partners
AI can extend thinking, but shouldn’t replace it. Our work avoids tool addiction and focuses on human judgment and shared language.
Who we work with
Most people who reach out already think deeply about their work. They’re not looking for magic AI; they’re looking for clarity, structure, and sustainable practice.
- Founders and early teams shaping products or research tools.
- Engineers and researchers designing long-lived systems.
- Teams introducing AI without losing their own language or judgment.
- Leaders dealing with complex decisions, transitions, or cross-role work.
- Organizations preferring steady structural change over big-bang transformation.
About Fragment Practice
Fragment Practice is led by Yasuhiro Shinsho, a thinking architect and full-stack practitioner based in Japan. The studio combines systems design, operations, and the craft of writing into a single practice centered on clarity and reasoning.
Our research backbone spans 2012–2025 across security, internal systems, AI adoption, document workflows, and personal knowledge systems. Fragment, Prism, and QFS grew from those long-term threads.
If you’re carrying a heavy document, a messy folder of notes, or an unnamed tension, you can start there. A single fragment is enough.
Related pages
Services
Design of AI usage, workflows, and quiet internal systems.
Research & Writing
Frameworks, structural notes, applied research, and ZINEs.
Fragment
A calm, AI-native notebook built on Markdown × YAML.
Contact
Start with a small fragment or question.