The operational library
How we actually work
Not a blog index: these are the real processes we run for clients and on our own business, written down step by step, with the tools, honest timelines, and the cases where each workflow is the wrong choice.
Build
Claude Code Skills: Our Agentic Engineering System
How we standardize engineering work as reusable Claude Code skills, so every build starts from a proven playbook instead of a blank prompt.
5 steps · 1-3 weeks depending on how many processes we capture as skills
AI Product Build: From Idea to Production
Our end-to-end process for shipping an AI product or feature: scope the risky part first, build on proven rails, launch with evals and guardrails.
5 steps · 3-8 weeks: one week to de-risk the AI core, the rest is shell, guardrails, and rollout
Automate
Grow
Lead Qualification Workflow
Capture, enrich, score, and route inbound leads automatically, so hot leads get a booking link in minutes and cold ones never eat a sales hour.
5 steps · 2-4 weeks depending on how many capture points and CRM quirks are involved
SEO + GEO Content System
The content system we run on our own site: research from real search data, human-reviewed drafts, schema, internal links, and a ledger where every change is a tracked bet.
6 steps · 2-4 weeks to install the system; rankings compound over months, and anyone promising faster is selling something
GTM Automation
Outreach research and personalization pipelines with hard safety rails: humans approve everything before it sends, volume stays capped, and your domain stays clean.
6 steps · 3-5 weeks to a working pipeline; expect another few weeks of approval-queue feedback before draft quality settles
Operate
Customer Support Automation
Automate support the safe way: triage everything, draft replies for human review first, deflect FAQs with grounded answers only, and earn autonomy with data.
5 steps · 3-6 weeks to reach reviewed drafts in production; autonomy decisions come after real usage data, not on a launch date
AI Agent Deployment
How we take an agent from working-on-a-laptop to production: scoped permissions, real secrets handling, sandboxing, eval gates, observability, and rollback.
6 steps · 2-6 weeks depending on how many systems the agent touches and the state of the existing code
Free weekly brief
Steal these workflows
Get new workflow teardowns from this library by email: steps, tools, honest timelines. No spam, unsubscribe anytime.