Building NiallOS with Claude Code
“AI will save you so much time!” Silicon Valley promised. So I tried everything. ChatGPT for writing assistance. Claude for research summaries. Gemini for project management. Copilot for planning. Every morning started the same way: Ten minutes explaining my context to whichever AI I was using that day.
“I’m a Product Lead at Cambridge Education Futures…”
“I’m working on a Master’s…”
“I need this in British English for an internal audience…”
“Keep it practical, I’m building training programmes…”
Then I’d finally get to the actual work. AI was supposed to save time. Instead, I felt busier than ever.
I was drowning in AI tool overload while still doing everything manually.
The AI Promise vs. Reality
2023-2024: AI will transform knowledge work! Automate everything! Work 4 hours a week!
That was the promise. My reality as a Product Lead at Cambridge looked quite different. I was designing training programmes for education leaders working with organisations like UNICEF and KFAS, and managing a Master’s research project. Creating workshop materials and training curricula. Publishing reports, papers and research outputs. Tracking professional development and learning goals. Staying current on EdTech and AI in education.
AI was supposed to save time on routine tasks, help me focus on strategic work, make complex work manageable, and give me breathing room.
What actually happened? Every morning became an AI context-switching marathon. ChatGPT for brainstorming workshop activities. Claude for analysing research papers. Back to ChatGPT for LinkedIn post drafts. Gemini for curriculum outline suggestions. Each tool required a fresh start and re-explaining everything from scratch.
I calculated the daily context tax. I wasn’t using AI to work smarter. I was managing a portfolio of AI tools, each one requiring setup, context, and translation. It felt like having five junior assistants who all had amnesia every morning.
The Breakthrough Moment
I was watching a podcast where a startup founder demonstrated how he’d set up his Claude Code life management system. He showed slash commands originally designed for software development workflows. /commit automatically generated git commits. /review-pr analysed code changes. Custom commands executed his specific project patterns.
But what caught my attention was how he’d adapted these development tools for managing his daily tasks. No re-explaining project structure. No describing preferences. The system just knew his contexts, his conventions, and his workflow.
That’s when it hit me: What if I could build slash commands for my WORK?
Claude Code offered something fundamentally different from generic AI chat interfaces.
First, access to your file system. Claude can read your actual project files, see your folder structure and organisation, understand context from existing documents. No manual copying and pasting required.
Second, access to your terminal. It can run commands on your behalf, execute workflows automatically, and integrate with your actual tools. Real automation, not just text generation.
Third, persistent customisation. You can create your own slash commands, build workflows specific to your needs, and the system remembers your preferences. It evolves with your work patterns.
The difference between generic AI chat and Claude Code became clear. Generic tools like ChatGPT and standard Claude have no access to your files. You might connect folders like Google Drive, but it’s still a manual process. No memory is retained between sessions (unless you activate this feature in your Pro account). Can’t execute workflows automatically. One-size-fits-all interface. You adapt to the tool.
Claude Code reads your actual files and folders. Maintains persistent custom commands. Executes complex workflows. Customised to your specific needs. The tool adapts to you.
Every AI tool I’d been using treated me like a stranger every single day. But Claude Code could access my actual work, see my patterns, remember my preferences, and execute my workflows the way I actually work.
This was a foundation for building a personalised AI operating system.
Building NiallOS: The Personal Operating System
I spent a weekend asking myself: If I could teach AI everything about how I work, what would I tell it?
Not one-off instructions. A comprehensive profile. A foundation.
What I built is NiallOS – my personal AI operating system, built on Claude Code. It’s not an app. It’s not a productivity hack. It’s an operating system for my professional life. It’s just folders and markdown files, so I could plug it into Gemini’s CLI or another LLM if needed. It also lives locally on my machine with constant backup.
The Digital Twin
I created a file called digital_niall.md as a comprehensive profile that every AI workflow reads before executing.
My professional identity includes my role as Product Lead at Cambridge Education Futures, focusing on AI training programmes for education leaders. My expertise spans EdTech, AI literacy, and learning science. Current projects include the UNICEF Innovation Camp, the KFAS training programme, and the Cambridge Digital Leadership Programme.
My academic context captures that I’m a Master’s student at UCT with a research focus on AI literacy development and transformative learning frameworks. My writing standards specify British English and APA formatting with rigorous page-numbered citations. Citation management follows strict academic integrity protocols.
Communication preferences are clear: a casual yet technical tone that’s professional yet accessible. Moderate length with enough detail but not overwhelming. Plain English over jargon. The primary audience is education leaders, practitioners, and researchers.
Work patterns matter too. Peak productivity hits between 6 and 9 AM. Energy drainers include excessive meetings (more than 4 per day) and repetitive admin tasks. Energy sources include workshop design, research synthesis, and new programme development. Time constraints are significant with a full-time role, a Master’s degree, and multiple projects.
My priorities follow a clear hierarchy: the Cambridge projects come first as a professional responsibility. Research quality for the Master’s degree completion is second. Professional development in AI, Python, and automation is the third priority. Time efficiency is fourth, given resource constraints and the need for leverage.
This digital twin knows everything about my professional context, my standards, my constraints, and my goals. Once embedded, every AI workflow reads this profile first. No more re-explaining. No more context tax.
Slash Commands
With my digital twin in place, I built custom workflows as slash commands.
Daily workflows include /daily-brief for personalised news on EdTech, AI in education, and policy developments. /daily-checkin handles evening reflection and learning tracking.
Weekly systems include/weekly-checkin for progress tracking across projects and learning goals, and /project-tracker for a strategic dashboard of all active work.
Academic and research commands include /research-manager for literature organisation, citation management, and gap analysis. /academic-writing produces British English scholarly writing drafts with proper citations. /create-citation-cards generates batch citation cards with page numbers.
Professional development tools feature /learning-path for structured learning roadmaps from Foundation to Mastery. /study-session tracks learning with knowledge checks. /learning-review provides weekly learning pattern analysis.
Content and training commands include /workshop-planner for comprehensive workshop design, including scripts and materials. /curriculum-planner handles full course design, complete with learning objectives. /rewrite-for-website transforms workshop notes into portfolio articles.
Specialised AI Agents
Each slash command launches specialised AI agents that read my digital twin first, then execute specific tasks. The metrics analyst analyses progress across projects. The research organiser manages academic literature. The workshop designer creates training materials. The learning path builder designs structured curricula. The content adapter transforms content for different formats.
The philosophy underlying all of this is that instead of adapting myself to generic AI tools, I built a system that adapts to me. Every workflow knows who I am, what I’m working on, and how I prefer to work.
Why Personalisation Changes Everything
The contrast between generic AI and personalised AI is stark.
Every generic AI conversation starts fresh. “I need help planning a workshop on AI literacy…” The AI asks, ‘What’s your audience?’ (education leaders). What level (executive decision-makers)? How long (90 minutes)? What’s your goal (practical understanding plus policy implications)? Five minutes of context-setting before getting actual help.
With personalised AI through NiallOS, I type /workshop-planner ai-literacy and the output already knows everything. The audience is education leaders and policymakers. Format is 10-10-10 (instruction, demo, practice). Style is practical, accessible, and story-driven. Duration is 90 minutes, my standard format. Standards include accessibility, Q&A prep, and troubleshooting. Zero context-setting. Straight to value.
The compounding effect is powerful.
One-time setup cost involves two to three hours building the digital twin, a few hours creating first slash commands, and ongoing refinement as you learn what you need.
Ongoing returns include every command, saving 15 to 45 minutes of context explanation. Quality improves because the system is aware of your standards. Consistency increases through the same framework every time. Customisation deepens as the system learns your patterns.
Who benefits most from personalised AI? High-value candidates include researchers managing complex literature and citations, educators creating training materials and curricula, programme managers juggling multiple projects, academics balancing teaching, research, and service, professionals with clear preferences and repeated workflows, and anyone who regularly finds themselves re-explaining context to AI.
The personalisation principle states clearly: Generic tools ask, “What do you want?” Personalised systems know “Who you are, what you’re working toward, how you work.”
The most effective productivity system is the one that understands you.
The Article Series Ahead
Over the following seven articles, I’ll show you exactly how NiallOS works and how you can build your own personalised AI operating system.
Article 2: The Digital Twin – Teaching AI Who You Are will explore what a digital twin is and why it’s the foundation, the 77 questions I answered to build my profile, how every workflow uses this context automatically, and how to build your own from minimum viable to comprehensive.
Article 3: System Refinement – How It Learns and Improves Itself explains why static productivity systems fail over time, how NiallOS analyses my work patterns and suggests improvements, real refinement sessions that caught major blindspots, and building self-improving systems.
Article 4: Project Management – Managing Complexity Without Chaos demonstrates how I track eight active projects without spreadsheets, the /project-tracker workflow in detail, energy management and preventing burnout, and strategic decision-making in minutes not days.
Article 5: Research Assistant – Academic Work at Scale covers managing Masters research and literature systematically, citation management and academic integrity, the /research-manager and /create-citation-cards workflow, and finding research gaps and synthesis insights.
Article 6: Workflow Automation – Slash Commands That Save Hours teaches building custom slash commands for your specific needs, daily and weekly routines automated, content creation workflows, and the compound effect of small automations.
Article 7: Learning Paths – Structured Skill Development with Proof shows how I learn new skills (Python, GenAI, automation) while working full-time, the /learning-path approach from Foundation to Application to Mastery, assessment and validation systems, and building verified competencies with portfolio evidence.
Article 8: The ROI of a Personal AI Operating System provides measurable outcomes across time, quality, learning, and wellbeing, the business case for building personalised AI, who should do this and how to start, and the future of AI-augmented knowledge work.
This isn’t about copying my system. It’s about understanding the principles of personalised AI so you can build your own operating system tailored to your work, your goals, your preferences.
What you’ll learn: How to teach AI your context persistently. How to build workflows that adapt to you, not vice versa. How to create self-improving productivity systems. How to leverage AI for genuine leverage, not just another tool to manage.
This is Article 1 of 8 in the series “Building Your AI Operating System with Claude Code”