01 / Strategy
Cut surface area, not coherence
Broad intelligence product, reduced to three two-week sprints.
PC Deals, comparison, Portfolio Benchmark, and Shared Requests.
Selected product decisions / 01—16
How I think when the pieces do not line up.
Problems, constraints, rejected alternatives, decisions, and honest trade-offs.
02 / Data model
One deal, one current view, history underneath
One deal can contain term sheets, preliminary reports, final agreements, and amendments.
Use the latest published report and representative benchmark tranche.
03 / Data model
Do not aggregate tranches when semantics are undefined
Tranches have different covenant semantics.
Show the representative benchmark tranche and its actual name.
04 / Data model
Separate identity from occurrence
Report-level TrancheIDs change even when the economic facility persists.
DealID → TrancheLineageID → report-level TrancheID occurrences.
05 / AI & trust
Automate candidate discovery before automating judgment
Borrower name alone cannot determine whether an agreement belongs to an existing deal.
Automate candidate discovery before judgment.
06 / AI & trust
Make human review become exception handling over time
Default human review limits scale, but unsupported autonomy creates risk.
Capture corrections, measure errors, calibrate, then automate high-confidence cases.
07 / AI & trust
Private data boundaries belong in the architecture
Private agreements are client-owned; global AI search can cross boundaries.
Client-scoped sources and anonymized market aggregates.
08 / Data model
Missing is not No
Missing values can silently become negative observations.
10 Yes / (10 Yes + 5 No) = 66.7%; exclude 5 Missing.
09 / Workflow
Put the hard gate at the consequential action
Dependent required fields initially blocked saving partial work.
Put the hard gate at the consequential action.
10 / Workflow
Complete for what?
A generic Complete flag collapses multiple readiness states.
Extraction complete; scoring 100%; report 87%; publication blocked.
11 / AI & trust
Professional judgment is part of the product model
A calculable score is not automatically a legitimate assessment.
Structured facts + deterministic factors + attorney judgment.
12 / AI & trust
Feedback loops require lineage by design
A reviewer correction is useless upstream if its origin is lost.
Track payload identity, source path, model version, original value, correction, evidence, and reviewer action.
13 / Workflow
Deterministic fact outranks probabilistic similarity
Exact duplicates and similar precedents answer different questions.
Deterministic fact outranks probabilistic similarity.
14 / Strategy
Prototype to create disagreement while it is cheap
Words hide disagreement about navigation, data density, and semantics.
V1 broad concept → V10 dashboard/screener/benchmark/requests → V14 narrower loop → V15 sharper screening.
15 / Workflow
Build the customer workflow, not the org chart
The workflow crosses portal, data factory, Studio, scoring, reports, and publishing.
Submit → Track → Receive → Analyze.
16 / Workflow
A successful API call is not a successful experience
A success message appeared without a usable score or destination.
Treat missing score or URL as an incomplete outcome.