Analyze opportunity data
Inspect grain, definitions, duplicates, missingness, and calculations before interpreting a dataset.
Data-quality and opportunity analysis
Finish with an artifact another professional can review, challenge, and act on.
When to use this
Inspect grain, definitions, duplicates, missingness, and calculations before interpreting a dataset. Use it when the decision matters enough that source authority, uncertainty, and human review must stay visible.
Match the depth to the consequence
Quick
Produce the minimum useful data-quality and opportunity analysis from the controlling sources and flag gaps.
Working
Compare sources, classify evidence, produce the data-quality and opportunity analysis, and run the task-specific checks.
Thorough
Add a countercase, conflict review, stakeholder challenge, and explicit validation plan before action.
Bring only what the task needs
- The decision this work must support
- Current, approved source material
- Known constraints and open questions
Boundary: Approved customer information in an employer-approved environment; otherwise use public or fictional material.
Do not add credentials, unnecessary personal information, private security details, unapproved pricing, contracts, or customer material.
Work sequence
- State the decision the data-quality and opportunity analysis must support.
- Select the controlling sources and record their date, owner, and authority.
- Separate customer evidence from seller interpretation before asking AI to synthesize.
- Build the data-quality and opportunity analysis and label Known, Inferred, Assumed, and Unknown.
- Challenge the result with contrary evidence and complete the verification checks.
Analyze opportunity data
GOAL Use approved structured data to support a decision without hiding quality problems. CONTEXT The account, audience, decision, and current sales stage are [CONTEXT]. CONTROLLING SOURCES Use only [APPROVED SOURCES]. Identify the source behind every material fact. TASK Restate the row grain and definitions, identify duplicate and missing records, show calculation logic, separate patterns from possible explanations, and name rows requiring review. CONSTRAINTS Do not invent customer intent, authority, dates, budgets, pricing, commitments, product capabilities, legal conclusions, or missing facts. REQUIRED OUTPUT Data-quality and opportunity analysis UNCERTAINTY BEHAVIOR Separate Known, Inferred, Assumed, and Unknown. Show conflicts instead of resolving them without evidence. Turn important unknowns into questions. VERIFICATION REQUEST List the claims, calculations, names, dates, owners, commitments, and source conflicts a human must verify before acting. Construct the strongest evidence-based case against the proposed conclusion.
Northstar example
The Northstar worksheet has one row per workstream, a duplicated workshop row, an undefined stage label, and renewal data that conflicts with procurement.
Task-specific checks
Why this works
The sequence starts with the decision and source authority, so fluent output cannot quietly replace evidence. The countercase turns uncertainty into focused customer questions.
A weak AI request
“Create a polished data-quality and opportunity analysis and make it persuasive.”
This request has no controlling sources, evidence boundary, uncertainty behavior, or verification step. It rewards confident prose instead of dependable work.
Practice safely
Find the duplicate, missing definition, and renewal conflict in the Northstar opportunity worksheet.
Open Northstar exercises