Private and public positions in one balance sheet, owned through the real entity structure, scoped to the branch that’s asking. Dashboard figures drill straight to the ledger.
GP notices, custodial statements and held-away accounts become structured ledger entries, reconciled to the source. Backfill brings in years of history at onboarding, and anything low-confidence waits in a review queue before it reaches the books.
One commitment ledger per position, return of capital applied first, and performance computed from the ledger itself rather than keyed in from GP-reported figures. Realizations get flagged from GP correspondence before the distribution hits.
“…the General Partner has entered into a definitive agreement for the sale of the Bayview portfolio, expected to close in Q1…”
Can the balance sheet absorb the next 18 months of calls without forced selling in the liquid sleeve? A reproducible pacing engine, a liquidity ladder and bad, base and good cases that run through the same model. No Monte Carlo noise, and no LLM in the numbers.
Query the consolidated book in plain language, cite the page a document answer came from, and produce the quarter-end pack from the same ledger. Entitlements are applied to the data before the model sees it.
Each family’s book segregated from every other. A house view for the MFO CIO across all of them. And the operating controls compliance asks for: batch maintenance, a who-did-what trail and AI cost attributed per family.