The setting

Northbridge sells cloud bookkeeping software to small firms. You are the marketing operations manager, keeping the CRM, lead flow, automation, and reporting clean for a four-person marketing team.

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. CRM auditstarter

Clean up the Northbridge lead export before the quarterly sync

Before the quarterly data sync to the sales system, the lead table needs a scrub. Right now it is full of duplicates, stale open leads that will never convert, and inconsistent source labels that break reporting. If this syncs as-is, sales inherits the mess and the numbers lie.

You deliver: A CRM audit listing duplicates to merge, stale leads to close, data errors, and source values to normalize, with lead ids and counts.

Scored on: Finds the duplicates, Finds stale leads, Catches data errors, Normalizes sources.

Working from: crm_leads.csv, data_standards.md.

2. segment logicstarter

Define the segment logic for Northbridge's year-end tax push

Northbridge wants to email existing customers about upgrading before tax season, when small firms feel the pain of manual books. A sloppy segment would email people who unsubscribed or already left, which is both a bad look and a compliance problem. You need airtight segment rules.

You deliver: A segment definition with inclusion and exclusion rules and the resulting contact list from the sample.

Scored on: Correct inclusions, Excludes opt-out and churned, Plan scope, Reusable rules.

Working from: field_dictionary.md, sample_contacts.csv.

3. workflow QAcore

QA the Northbridge lead nurture workflow before it goes live

A new five-email nurture workflow is scheduled to go live tomorrow. The last one had an embarrassing bug that emailed unsubscribed people, so this time it gets a real QA pass first. You have the spec and a log of what test contacts actually received.

You deliver: A QA report listing each workflow defect with evidence, severity, a fix, and a go or no-go recommendation.

Scored on: Missing conversion exit, Missing opt-in check, Broken greeting token, Severity and recommendation.

Working from: workflow_spec.md, test_log.csv.

4. attribution checkcore

Reconcile the double-counted pipeline in Northbridge's attribution report

The marketing attribution report claims 136000 dollars of influenced revenue this quarter, but finance closed only 50000. Leadership is about to reallocate budget based on the report, and the numbers do not add up. The report is crediting full revenue to every channel that touched a deal, counting the same money two...

You deliver: A reconciliation note explaining the double-count, the true total, the multi-counted opportunities, and corrected per-channel credit under a sound model.

Scored on: Explains the double-count, True closed total, Identifies multi-counted opps, Corrected model sums right.

Working from: attribution_touches.csv, summary_report.md.

5. handoff notestretch

Write the ops handoff note before Northbridge leave

You are out for two weeks starting Monday and a junior teammate is covering marketing ops for the first time. If the wrong things break while you are gone, campaigns stall and data gets worse. The handoff has to make the risks and the in-flight work impossible to miss, ranked by what actually matters.

You deliver: A handoff note covering systems and owners, risk-ranked open issues, in-flight work, and the critical items to watch.

Scored on: Surfaces the critical sync, Flags the compliance risk, Risk-ranked issues, Owners and actions.

Working from: systems_inventory.md, open_issues.csv.

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 marketing operations" something an employer can check.