Claude Code is now our most-used interface across 20 team members..
The way we run Workflows today wouldn’t have been possible a year ago. We bet big on something we call AI-Native Services. It pairs software and human expertise inside one operating layer.
Here’s exactly how it works:
1. Company OS in Git
Our ENTIRE company dataset lives in one GitHub repo called Company OS.
What’s inside:
company/: team, voice guide, design system, industry intel
wiki/: SOPs, playbooks, campaign guides
clients/: per-client context files
raw/: client calls, market research, competitor data
plugin/: 26 agents, 23 commands, hooks
skills/: 79 Claude skills
Data flows in constantly to keep it up-to-date.
2. Client repos
Every client gets their own private repo.
Same engineering pattern as the Company OS, just personalized to their account.
What's inside:
their ICP, voice guide, and brand assets
historical campaigns and what worked
onboarding form data and deep research
Slack threads, call transcripts, GDrive changes
API and MCP connections to their revenue stack
Every team member walks into a session with full client and company context already loaded.
3. Human interaction layer
We still log into some SaaS UIs, but Claude is slowly taking over.
Across 20 team members, the efficiency gain has been massive. We automate as much of the admin layer as possible:
client onboarding
content research and ideation
skill tuning from team feedback
reply triage and sentiment routing
campaign launch pre-flight checks
AI does the legwork. Humans ship.
We get more time for strategy and creative GTM.
4. MCP + CLI engine
MCPs and CLIs let Claude act across our stack instead of just advising.
Some of our favourites:
GitHub: Company OS + client repos
Findymail: email verification waterfall
Google Workspace: client docs
Airtable: automation backend
Instantly.ai: email campaigns
Slack: team and client comms
Apollo.io: list + enrichment
Notion: internal wiki + PJM
HeyReach: DM sequences
Plus HubSpot, Browserbase, Supabase, Vercel, Figma, Stripe, Pinecone, Clay, Apify, Firecrawl, and more.
We're also migrating more workflows to custom code.
5. Operationalize
We're constantly optimizing how we work with AI and software. Built into the system:
Guardrails: Safety hooks gate 94+ risky operations.
PR-based governance: Anyone on the team can propose a new skill, agent, or tweak as a branch.
Workflows-engineering plugin: 26 agents, 79 skills, 23 commands auto-propagated. Agent swarms split tasks into 5-20 sub-agents.
Self-improvement loop: n8n syncs tech stack data back into the Company OS. Pinecone stores past content and performance metrics for skills to query. Human corrections feed back in.
There's no finish line. We're building like it's a marathon.
So what do you do with this?
There’s no shortcut to this. If you want AI to run your operations, the operating layer has to come first.
Start with one shared repo for your team’s context. Wire in agents, MCPs, and skills from there.
We built ours from scratch and run it every day. If you want help figuring out what an AI-native version of your business looks like…
Book a meeting here.
We'll walk you through it.