OnTracAI

Opportunity Identification (Raw Observations)

Purpose

This is the raw material layer.

Opportunity Identification captures direct signal from interviews and kickoff discussions before any synthesis, scoring, or prioritization happens.

Nothing enters this layer unless it is directly attributable to something that was said.

This is where:

  • Quotes are preserved
  • Nuance is retained
  • Premature abstraction is avoided

Canonical Opportunity Log

This is maintained as one master sheet across all functions.

Required Columns

ColumnDescription
FunctionIT, Finance, Ops, Sales, etc
ObservationFactual statement of what exists
TypeIssue / Strength / Idea / Constraint
ImpactWhy it matters
Example or QuoteVerbatim or close paraphrase

Every row must tie back to:

  • A transcript
  • The kickoff
  • A system walkthrough

If it can't be sourced, it doesn't belong here.


Writing Rules

RuleGuidance
Be concreteAvoid abstractions
Separate fact from opinionObservation is not interpretation
One idea per rowNo bundling
Neutral toneNo judgment
Preserve languagePrefer exact wording

Good observations feel boring. That's intentional.


Example

FunctionObservationTypeImpactExample or Quote
ITThree ERP systems maintained in parallelIssueIncreases integration complexity"We operate on three ERP systems…"
ITSnowflake DW with CDC and DBTStrengthEnables unified analytics"The glue that keeps them together…"
ITSnowflake access restricted to ITIssueCreates analysis bottlenecks"No one outside IT can query it."
ITConsidering Sigma for self-serviceIdeaCould reduce IT load"Sigma looks and feels like Excel."

This pattern repeats across every function.


Allowed Observation Types

TypeMeaning
IssueActively causing pain or drag
StrengthExisting asset to build on
IdeaSuggestion or exploration
ConstraintHard boundary or policy
ObservationNeutral fact without judgment

Use "Observation" when unsure. Interpretation comes later.


Attribution Discipline

SourceHandling
Direct quoteUse quotation marks
ParaphraseStay close to wording
Multiple interviewsNote consistency later
Single mentionStill capture it

Attribution builds trust in the final readout.


How This Feeds Synthesis

This log is later used to:

  • Deduplicate similar issues
  • Identify shared root causes
  • Surface cross-functional enablers
  • Separate symptoms from causes

No scoring happens here.

No prioritization happens here.

This is evidence collection.


Success Signal

This step is successful if:

  • The sheet feels long and unglamorous
  • Multiple rows say similar things in different words
  • Quotes jump off the page

That redundancy is the signal.

Next step is Scoring & Prioritization, then Theming, where these raw cuts collapse into something leadership can digest.