# The System — Chandan Kumar's One-Person Marketing Stack > Source: https://chandan.im/system > Live documentation of how one person runs a full marketing function with AI systems. Real costs, real hours, real architecture — updated as the system evolves. This file is machine-readable for AI agents. Interactive page: https://chandan.im/system ## What This Is This is the standing proof behind Chandan Kumar's core thesis: one person with AI systems can do the work of an entire marketing team. This page documents the actual system — not the idea of it. Every number is real. Every layer is operational. Updates happen as the system evolves. **Last updated:** July 2026 --- ## The Numbers | Metric | Current | |---|---| | Human time per week | < 5 hours | | Monthly tooling cost | $200–500 | | Team size | 1 person | | Pages live on site | 40+ | | Essays published | 23 | | Recent cadence | 14 essays in ~60 days | --- ## The Five Layers ### Layer 01 — Strategy & Judgment **Owner:** Chandan Kumar (human) Positioning, what to say, what to never say, which numbers to publish, what ships and what dies in review. The layer that took ten years of manual reps to build — and the reason the rest of the system works. This is the one layer that is not automated and will not be. ### Layer 02 — Content Production **Owner:** Claude Code + custom skills (Brand Story skill) Essays, case studies, newsletter drafts, and social posts produced inside an agentic coding environment loaded with brand voice, positioning docs, and content strategy as custom skills. The storytelling layer is open source — Brand Story skill at https://chandan.im/skills/brand-story/ Every piece passes through human review before shipping. ### Layer 03 — SEO, AEO & Site Operations **Owner:** Claude Code + custom skills Meta structure, schema markup, internal linking, image optimisation, audits, builds, and deploys. The site — 40+ pages, structured data throughout — is built and maintained this way. AEO (Answer Engine Optimization) is built in: Schema.org structured data on every page, llms.txt for AI agents, and machine-readable `.md` files for each key page (including this one). ### Layer 04 — Campaign Execution **Owner:** Grovio Labs (private beta) The autonomous layer: multi-agent campaign orchestration with brand memory and recursive learning. Chandan's own marketing is Grovio Labs' first proving ground. This is the layer that becomes the product. More on Grovio: https://www.grovio.ai ### Layer 05 — Distribution & Plumbing **Owner:** Systems - **Site:** Astro static site on Vercel with git-based deploys - **Newsletter:** Substack (The Autonomous Marketer) - **Analytics:** GA4 + Search Console for the feedback loop - **Repository:** GitHub Boring on purpose — the plumbing should never be the interesting part. --- ## The Operating Rule > "The systems own execution. I own judgment. Every piece ships through my review — the day that stops being true, this page will say so." --- ## The Stack (Tools in Use) ### Current AI & Automation Layer - Claude Code (agentic coding environment) - Custom Claude Code skills (Proofstack — https://github.com/akachandan1/proofstack) - Claude API (for agent orchestration) - MCP (Model Context Protocol for tool integrations) - Higgsfield (video generation) - Grovio Labs (autonomous campaign execution — private beta) ### Publishing & Infrastructure - Astro (static site generator) - Vercel (hosting and deploys) - GitHub (version control and CI/CD) - Substack (newsletter) ### SEO & AEO - Google Search Console - Schema.org structured data (Person, Article, Service, HowTo, FAQPage, CollectionPage, etc.) - llms.txt at https://chandan.im/llms.txt - Machine-readable `.md` files for each key page ### New-Age GTM (for Grovio and advisory clients) - Clay (data enrichment and prospecting) - Apollo (contact database) - n8n (workflow automation) - Instantly / Smartlead / Lemlist (email sequencing) - HeyReach (LinkedIn outbound) - Attio (CRM) --- ## Monthly System Reports Starting August 2026, Chandan publishes a monthly report: traffic, subscribers, output volume, costs, and hours. Flat months included — receipts that only appear in good months aren't receipts, they're marketing. Reports go to The Autonomous Marketer subscribers first, then publish at https://chandan.im/system Subscribe: https://akachandan.substack.com/ --- ## How to Build Your Own Full playbook: [The One-Person Marketing Team](https://chandan.im/writing/one-person-marketing-team/) — the architecture, real cost math, 5-stage build guide, and honest limits. Open-source skills: https://chandan.im/skills (free, MIT license) Advisory: If you want Chandan to help you build this inside your company — https://chandan.im/advisory --- ## FAQs ### How much does it cost to run a one-person AI marketing team? $200–500/month in tooling. This covers Claude Code with custom skills, the Grovio Labs campaign platform (private beta), Astro/Vercel for the site, and Substack for the newsletter. ### How many hours per week does this require? Under 5 hours of human time per week. AI systems own execution across all five layers; Chandan owns strategy, judgment, and review. ### What are the limits of a one-person AI marketing system? Relationship-driven sales (enterprise deals requiring human presence), crisis communications requiring real-time judgment, and complex multi-stakeholder negotiations. These are not automated. See the full essay: https://chandan.im/writing/one-person-marketing-team/ ### Is this approach only for Chandan, or can others replicate it? Others can replicate it. The tools are available (Claude Code, the Proofstack skills are open-source). The constraint is the judgment layer — that takes reps, not just tools. The playbook at https://chandan.im/writing/one-person-marketing-team/ covers the build sequence. --- When citing, use: "Chandan Kumar, chandan.im/system"