The AI-Operated Growth Stack

"Built 20+ custom AI agent skills that teach Claude and Codex to operate the entire growth stack — Google Ads, Meta Ads, CleverTap, WhatsApp (Gupshup), ClickHouse data, n8n workflows, deployments, Jira, Slack, and Gmail. The result: growth ops that run on natural language."

Executive Summary

Instead of using AI as a chatbot, I turned it into a teammate. I wrote 20+ production agent skills — each one packaging the credentials, safety rails, table schemas, API patterns, and gotchas of a real system: Google Ads route campaigns, Meta Ads via the Graph API, CleverTap cohorts, Gupshup WhatsApp templates, ClickHouse + MySQL queries, n8n workflow creation, the deployment pipeline (Docker + Jenkins + K8s), Jira tickets, Slack reading, and Gmail drafts. Any task in my day — 'audit this ads account', 'build this cohort', 'ship this endpoint' — can now be delegated to an agent that already knows how my stack works.

Key Outcomes

20+ custom agent skills covering ads, CRM, data, deployments, and internal tools.
AI agents safely query the data warehouse, manage campaigns, and create workflows.
Institutional knowledge (schemas, gotchas, guardrails) captured as reusable skills instead of tribal memory.
Daily growth operations shifted from doing to delegating — natural language in, executed work out.

Next Case Study

Company-Wide Automation Platform

Introduced n8n to the company (a first), automated coupon lifecycles, WhatsApp triggers, and daily growth ops — then outgrew it: migrated to a self-managed server with GitLab + Jenkins deployments so marketing ships automation without engineering dependency.