The setting

Fernhaven ships curated houseplant boxes on a monthly subscription. Support runs through a shared inbox plus chat, with a three-person team covering 400 to 600 tickets a week.

Every mission on this path happens at the same company, so context carries over the way it does in a real job: the data you cleaned in mission two is the data the finance lead questions in mission four.

The missions

1. ticket triagestarter

Triage Monday's ticket backlog at Fernhaven

You run support operations at Fernhaven, a houseplant subscription company. It is Monday morning and 24 tickets came in over the weekend while the team was off.

You deliver: A triage plan that orders the backlog, counts categories, flags supervisor cases, and drafts two first-response templates.

Scored on: Priority ordering, Accurate category counts, Supervisor flags, Usable templates.

Working from: tickets_monday.csv, support_policies.md.

2. knowledge articlestarter

Write the repotting help article customers keep asking for

Repotting questions are Fernhaven's single biggest care topic: 61 tickets last month, and the care card only covers watering.

You deliver: A help-center article on repotting that answers the top customer questions using only the gardener's notes.

Scored on: Grounded in the notes, Answers the real questions, Structure and tone, Honest about gaps.

Working from: repotting_notes.md, top_questions.md.

3. escalation notecore

Escalate the Hartley account before it becomes a chargeback

Ticket T-4413 is now a supervisor case: Dana Hartley, a 192 dollar per quarter customer, has received three damaged boxes in a row, mentioned a journalist friend, and this is their third contact.

You deliver: A supervisor escalation note with timeline, account value, risk assessment, and a specific recommended remedy.

Scored on: Accurate timeline, Account value math, Risk assessment, Specific remedy.

Working from: hartley_history.md, hartley_account.md.

4. QA reviewcore

Score two agent conversations against the QA rubric

Fernhaven reviews four random conversations per agent each month against the QA rubric in `qa_rubric.md`. This month it is your turn to review.

You deliver: A QA review scoring both transcripts line by line with quoted evidence and specific coaching for each agent.

Scored on: Evidence-based scoring, Catches the real errors, Balanced review, Actionable coaching.

Working from: qa_rubric.md, transcript_marco.md, transcript_sam.md.

5. insight briefstretch

Turn a month of tickets into the ops changes Fernhaven needs

Leadership asked one question at the ops review: what is support seeing, and what should we change? You have the monthly ticket summary in `march_summary.csv` and the agent notes in `agent_observations.md`.

You deliver: An insight brief for leadership: the trends, the root causes, and three evidence-backed recommendations ranked by impact.

Scored on: Correct trend analysis, Root causes, Cost quantified, Ranked recommendations.

Working from: march_summary.csv, agent_observations.md.

How the scoring works

Each deliverable is graded against the rubric written for that mission. Separately, every mission on every path is graded on how you used AI, against the same four criteria:

  • Understood the task. The learner framed the goal for the assistant clearly instead of pasting the brief and hoping.
  • Grounded in the material. The learner directed the assistant into the provided files and based the work on them, not on invented facts.
  • Verified the output. The learner checked claims, numbers, or coverage against the source material before submitting.
  • Iterated with judgment. The learner refined weak parts of the draft with specific follow-ups rather than accepting the first answer.

Both scores, with the work behind them, go on your proof profile. That is what makes a claim like "I can use AI for customer support operations" something an employer can check.