Aegis
InterOpera

Meridian Fixed Income — Compliance Report

Every figure is computed by a deterministic engine, traced through a knowledge graph to its source, and reconciled to the firm's answer key.

A fund holds a portfolio of bonds and must follow a rulebook (allocation limits, risk caps, liquidity floors). This is the report card: for every rule, is the fund inside its limit or breaking it — and can every number be proven?

How to read a row

  • Value — what the fund actually is (e.g. 9% in high-yield bonds).
  • Limit — what the rule allows (e.g. max 15%).
  • Utilization — how much of the limit is used. Firm A shows % (60.0%); Firm B shows basis points (6000 bps = 60%).
  • Status — the verdict. OK within limit · AT LIMIT exactly at it · BREACH outside it.
  • vs key — matches the official answer key.

Click any row

to see its proof: the path through the knowledge graph, the exact source passage in the guidelines PDF, the difference vs the answer key, and which firm rule produced it.

Why it's trustworthy (the 5 guarantees)

  • Reproducible — same inputs always give the same numbers.
  • Traceable — every number links back to the rulebook passage it came from.
  • No AI numbers — figures come from a deterministic engine; the AI only writes the summary, and a firewall blocks any number it invents.
  • Matches the answer key — reconciles exactly to the expected report.
  • Firm-switchable — flip Firm A→B and 3 numbers change, by configuration only (no code change).

Try it: switch Firm A → Firm Babove and watch “Aggregate non-IG exposure” and “Largest GRE issuer” flip to BREACH — Firm B counts them differently.