agents_webinar_demos

This repository addresses the challenge faced by marketers (such as growth marketers and data analysts) in quickly translating data insights into action due to a lack of engineering support in data analysis, conversion rate optimization (CRO), and campaign management. It provides six ready-to-run AI agent demos covering core marketing scenarios, from funnel analysis and anomaly detection to landing page audits.

By offering complete prompt sequences, synthetic datasets, and example outputs, users can reproduce these workflows directly in AI coding tools like Claude Code, Codex, or Gemini CLI. Each demo follows an "insight → hypothesis → action" framework, guiding the agent to query data, generate hypotheses, and output executable code or reports.

This resource is ideal for marketing specialists, e-commerce operators, and data analysts looking to leverage AI for automating repetitive analysis tasks, generating test hypotheses, or building marketing automation pipelines. For product managers, it also serves as an excellent case study for understanding how AI can empower business decisions.

Users are advised to first read PLAYBOOK.md to understand the strategic context and framework, then start with demos that have clear inputs and outputs, such as Demo 2 (paid media audience creation) or Demo 6 (website visual audit). All data is synthetic, allowing safe, full-process experimentation without connecting to real business systems.