02 · Compare the artifacts
Supported work. Visible uncertainty.
This fictional training case uses fabricated company and contact records. It contains no real personal or confidential data and does not provide privacy, legal, employment, or technology-vendor advice.
OrbitWorks CRM repair
OrbitWorks is a fictional business-to-business maintenance software company with 14 sales and service users. Its CRM holds 8,420 company records, 11,780 contacts, and 1,360 open opportunities. An audit sample finds 23 percent of open opportunities have no dated next step, 18 percent lack a decision-role contact, and 11 percent are assigned to former employees. Two pipelines use different meanings for qualified, yet both feed one forecast. Sales moves opportunities forward when a meeting is scheduled; customer success uses the stage only after a written problem and buying process are recorded. A marketing import of 2,100 contacts is waiting. The source contains 430 probable duplicates, 280 records without a documented marketing basis, 17 personal notes copied into a free-text field, and country values in nine formats. The proposed automation sends four emails and creates a call task, but it has no stop rule after opt-out, reply, disqualification, or opportunity creation. A dashboard reports a 72 percent win rate because it excludes closed-lost records older than 30 days. A regional manager requests administrator rights for all representatives to fix records faster. The CRM also has three integrations whose owners left the company; one export token has not been rotated in 19 months. Leadership wants a clean forecast by Friday and suggests overwriting all records from the import. Learners must design a controlled repair that improves lifecycle clarity, data standards, follow-up, stage evidence, reporting, permissions, and change management. They must minimize unnecessary personal data, preserve auditability, and avoid claiming that a configuration alone establishes lawful processing or accurate forecasts.
Supported example — reference only
- Step or milestone
- Stop sequence on recipient reply
- Trigger or date
- Any recorded reply event; event schema and timing not supplied.
- Evidence and source ID
- M03-I01 (F09, confirmed: no reply stop condition)
- Owner
- CRM process owner and data steward
- Gate or threshold
- Sequence must stop before another automated outreach action and create an owner review task.
- Dependency or gap
- Reply event mapping, task due rule, test result, monitoring, and activation approval: not supplied.
- Status
- Learner-proposed control — production blocked
A well-handled evidence gap
- Step or milestone
- Integration-owner-loss exception
- Trigger or date
- Integration owner departs or credential exceeds approved age.
- Evidence and source ID
- M03-I02 (F11, confirmed)
- Owner
- Authorized system owner
- Gate or threshold
- Automation pauses or routes safely until ownership and credential review are complete.
- Dependency or gap
- Integration identities, owner assignments, credential policy, token rotation, and test evidence: not supplied.
- Status
- Security and ownership gap open
Flawed approach — do not copy
Marking this automation specification “approved and complete” without the required evidence or reviewer is a flawed submission. Stop production activation if any consent, suppression, reply, opt-out, disqualification, opportunity, error, or owner-loss path lacks a passing test.
Repair: Rework the automation specification as an evidence-backed draft, not an approved result. Define the automation entry event, eligible population, owner, purpose, and data prerequisites. List reply, opt-out, disqualification, opportunity creation, error, and owner-loss as mandatory stop or exception conditions. Specify task creation, due-date logic, monitoring queue, and fallback for every stop condition. Check the revision against this requirement: All required stop and exception conditions appear in the specification. If the required evidence is still absent, keep the decision blocked and identify the missing input or authorized reviewer.
Full case record, ambiguities and all assignments →