A synthetic version of the kind of customer work we actually do.
This is an application exercise built from the kinds of workflow, data, and handoff problems Goose Group keeps finding in the field. It gives you a concrete look at how we work, and us a concrete look at how you think.
We care about shared beliefs in practice: capabilities over systems, useful not done, factory thinking, and respect for how work actually gets done.
The Work
Goose Group works with teams whose competitive advantage lives in the judgment of their people. We help them capture what their best people know and make it operational. The goal is to change how work gets done and leave behind capability that stays.
That means the real work is: understanding a customer quickly, hearing what is actually blocked, deciding what should stay human, deciding where heuristics are enough, deciding where AI helps, making one useful move, and leaving behind artifacts that compound.
We are a small, senior team. Everyone has the same job: find ways to use my talent to help customers. That means something different every week.
Who tends to thrive
Think of the work as three overlapping domains: technical systems, business dynamics, and facilitation. Nobody is equally strong in all three. The baseline is fluency across at least two and curiosity about the third.
The strongest people we have worked with tend to have a history of building things for themselves to create leverage. They can sit with a customer and understand what is really blocking them. They can move between customer conversation, product judgment, workflow design, data modeling, implementation, and communication without getting lost. They use AI to give themselves more capability, not more theater.
The Engagement
The customer is a company called Goldfish Express. They run aquarium retail, live delivery with narrow timing and temperature constraints, and recurring office maintenance. Growth has strained a system built on whiteboards, spreadsheets, Slack, and tribal knowledge.
The engagement materials include a discovery transcript, a workflow design review, business context, ops notes, operational datasets, rules, and reference material. Some are public on this site. The rest are available through the MCP so the work stays structured instead of scattered across loose docs and links. You are expected to decide what matters and ignore what does not.
What to do: Understand the customer. Find the real friction. Make one useful move. Leave behind something durable that makes your reasoning legible.
What to send back: A repo with your work, plus a note on what you did, why, and what you would do next. A deployed URL and video walkthrough are welcome but not required. Include at least one durable artifact that shows how you think.
If the repo is private, invite @itsmikejoyce and @alexfinnemore.
Full submission guidance and the engagement summary are documented at /api/endpoints. The MCP is the working surface for the rest of the material.
Start
If this resonates, send your details and we will send back an API key and MCP endpoint so you can work through the engagement materials and submit your work.
One question we care about up front: What have you built for yourself to give yourself leverage?