Technical Work
Explore the customer-operations intake demo, owner views, SQL model, and Power BI files. These separate learning artifacts show the supporting technical work; they are not a fifth case study or a deployed production system.
Read Before We Commit →One request, from missing information to review.
Use the sample intake to see how I would check a handoff. The owner dashboard, SQL model, and optional Salesforce mapping below are retained learning artifacts. They do not resolve commitment 388763 or verify its quantity, specification, or timing.
Two separate examples: the owner dashboard summarizes December 31, 2025 work. Riverside is a historical readiness review dated September 10, 2025. It illustrates the missing-confirmation problem; it does not establish that the same issue remained unresolved in December.
What would make this request ready for its next owner?
- Request arrives. Select Westbridge’s sample commitment-readiness request. The account is active, but that alone is not customer confirmation.
- Identify the blocker. Leave customer confirmation unchecked and try saving. The request stays outside the ready queue. The historical Riverside review below explains why a location-level answer matters.
- Preserve the evidence. Select the location, record confirmation, and attach the source before saving. These are sample checkboxes; a real workflow needs the underlying record and confirmation date.
- Assign the next step. Save to see the assigned queue and follow-up target. An individual owner must still accept the work. Submit again to test a possible duplicate.
Fictional, browser-local demonstration. “Ready for owner” means ready for review, not approved for release. Saving does not update the source data or connect to a CRM.
Local demo records (0)
Saved in this browser across refreshes, separately from the 44 source requests and 120 source actions.
No sample requests saved yet.
SUPPORTING PROTOTYPEExplore the owner dashboard and connected recordsSeparate December snapshot · six owners · 120 actions
Noah Brooks
All assigned accounts · featured account A006A written review prompt based on this featured account’s visible condition. The prototype does not generate a question for every action.
Westbridge Group
Source Account ID · A006- ARR
- $150K
- Modeled risk tier
- High
- Adoption change
- −0.6 pts
- Open or escalated cases
- 0
Block Ready status without an account, request type, applicable location, customer confirmation, and source evidence.
Implemented for local demo records: the same account, sample location, type, and received date is flagged before saving. A production workflow would review possible matches rather than assume they are the same issue.
The demo saves a queue and follow-up target with each accepted sample request. External task creation remains a proposed implementation.
Track owner and status changes while preserving source evidence, follow-up work, and final resolution.
OPTIONAL CRM IMPLEMENTATIONHow this workflow could be configured in SalesforceObject mapping, validation, routing, Flow, reports, and acceptance tests
The workspace above is a conceptual customer-operations prototype, not Salesforce Lightning and not evidence of a deployed org. The supporting specification maps the same workflow to standard and custom Salesforce objects as one possible implementation path.
View optional Salesforce specification ↗The workflow totals reconcile to the SQL and Power BI model.
The workspace, SQL model, and Power BI files share source IDs and record definitions. North Star is a standalone workbook with different account IDs, so only its published totals and five named queue accounts are reconciled here—not presented as row-for-row lineage.
The supplied fictional risk score creates the review group; it does not predict churn or lost revenue. Workspace ownership and follow-up views use the same 120 action rows as the SQL and Power BI files.
RELATIONAL ARCHITECTUREInspect the underlying data modelDimensions, activity tables, relationships, and record grain
Why I kept the tables separate
Accounts, locations, cases, actions, and decisions do not represent the same kind of record. Keeping them in related tables—and joining them only where the IDs match—lets the report connect the work without inflating counts or ARR.
FILES + QAInspect the build files and reconciliation checksSQL, Power BI, DAX, source data, and complete build pack
How I built and checked the model
Schema and analytical queries
Primary and foreign keys, constraints, indexes, joins, CTEs, conditional aggregation, latest-record logic, and window functions.
View SQL queries ↗Complete three-page project
The cleaned .pbip project includes the semantic model, desktop and mobile layouts, source workbook, DAX measures, and opening instructions.
Download Power BI project ↓Download source workbook ↓Download DAX measures ↓Visible reconciliation checks
60 accounts, 22 accounts in the modeled High-risk group, $4.149M in group ARR, a $2.358M top-five subset, 470 cases, 11 overdue actions, and 44 decisions all reconcile.
Download model + build pack ↓The Power BI Project (.pbip), source workbook, and setup notes show the supporting model. These are learning-project artifacts, not production implementation or measured business results.