PM & AI Workflow Case Study

A Concept Case Study: Orchestrating AI Tools Across a High-Stakes Live Fan Experience

Type

Concept Case Study

Tool

Anthropic – Claude Sonnet 4.6

Google – Gemini 3.1 Pro

Year

2026

Description

Play With Eclipse: A Canopy Fanmeet is a concept case study documenting how a two-model AI stack was orchestrated across a live, high-stakes fan event. This is a concept case study. All events, characters, and organizations are fictional.

Eclipse x Canopy Fanmeet promotional image

Executive Summary

Play With Eclipse: A Canopy Fanmeet was a live interactive fan event in which 10 raffle-selected members of the Lunaire fandom played Canopy alongside all 5 members of Eclipse, streaming publicly on Kronos Live. This case study documents the full project lifecycle across four phases: creative strategy, large-scale data auditing, incident response, and community recovery. A two-model AI stack was used throughout, with Gemini handling data processing and sentiment analysis and Claude handling technical problem-solving and communications. When a hardware validation failure caused the Round 2 session to crash live, the response required all three layers of the workflow: Gemini’s audit trail to trace the failure, Claude’s reasoning to fix it, and the Product or Project Manager’s (PM) judgment to decide what happened next. The case study argues that AI does not replace the PM in high-stakes environments. It surfaces the decisions that only a person can make.

Anthropic Claude logo

Claude

The Problem Solver

Google gemini logo

Gemini

The Data Auditor

PM decision maker icon

PM

The Decision Maker

A mockup of the Eclipse lightstick, "Eclipse Ring"

AI accelerates the work, but it can slip up. A human still has to make the calls that matter.