The setting

Tannerhill Coffee Roasters supplies wholesale beans to about 140 cafes across Ohio from a single roastery in Dayton. You are the operations manager, responsible for everything between the roaster and the loading dock.

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. SOP draftstarter

Write the packing and dispatch SOP Tannerhill never had

Tannerhill has grown from 40 cafe accounts to 140 in two years, and the packing floor still runs on habit. There is no written procedure: Rosa trained the last two hires by demonstration, and each of them now does the steps in a different order. Last month two orders shipped with wrong roast dates and a big account...

You deliver: A packing and dispatch SOP with numbered steps, owners, quality checkpoints, and fixes for the four observed gaps.

Scored on: Complete step sequence, Closes the four gaps, Checkpoints with criteria, Usable on the floor.

Working from: walkthrough_notes.md.

2. process mapstarter

Map order-to-dock at Tannerhill before the second shift starts

Tannerhill is adding a second packing shift in August, and nobody can hand the new shift a picture of how an order actually travels from the ShopVine website to the loading dock. You interviewed the three people who touch every order and their stories do not quite line up.

You deliver: A written process map of order-to-dock with numbered steps, owners, decision points, and a handoff-problems list.

Scored on: Complete end-to-end flow, Finds the double entry, Surfaces the contradictions, Rework loop shown.

Working from: process_interviews.md.

3. control checkcore

Run the June inventory control check on the green coffee counts

Tannerhill carries about 60,000 dollars of green coffee at any time, and the monthly count is the only control between the books and the warehouse floor. Your insurer asked for evidence the control policy is actually followed, and the June count sheet just landed on your desk.

You deliver: A control check memo listing every policy exception by lot id, with recomputed variances and specific remediation.

Scored on: Recomputes the variances, Finds the note exceptions, Segregation of duties, Specific remediation.

Working from: count_policy.md, june_count.csv.

4. RCAcore

Get past the obvious answer on the June 3 decaf incident

On June 3, twenty-seven cafes received decaf labeled as House Espresso, and Tannerhill spent 1,840 dollars on replacement roasts, overnight shipping, and credits. At standup the room settled on an easy story: the temp packer grabbed from the wrong bin. Elaine, the owner, does not buy it and asked you for a real root...

You deliver: A root cause analysis with an accurate timeline, a causal chain past operator error, and owned corrective actions.

Scored on: Chain past the operator, Finds the planted causes, Owned corrective actions, Accurate incident facts.

Working from: incident_timeline.md, floor_change_notes.md.

5. automation ideastretch

Pitch the automation that buys Tannerhill nine hours a week

Elaine has budgeted 4,000 dollars this year for one operations improvement and asked you to bring her a single automation proposal, not a wish list. You logged two weeks of manual task time and collected the surrounding facts to size the options honestly.

You deliver: A one-page automation proposal with a chosen target, ROI math from the logged data, an implementation sketch, and risks.

Scored on: Right target chosen, ROI math checks out, Plan grounded in the notes, Risks and manual keeps.

Working from: task_time_log.csv, context_notes.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 operations excellence" something an employer can check.