AI Credit Analysis demo
Ask a credit question in plain words; agents read the credit agreement as amended, the indentures, filings in their original language, call transcripts and your own notes, build the capital structure, and pin every cited sentence of the answer to the tranche it concerns. 1:34
What the demo shows
Answers credit analysts' questions across filings, indentures, calls and notes, all cited
Eight agents work each question. A Wall agent decides what the person may read before anything is chosen; a Planner splits the question and waits for the analyst to run the plan. An Indenture agent reads credit agreements as amended, indentures and intercreditors and computes the baskets; a Filings agent reads 10-Qs and foreign annual reports in the original language; a Call agent quotes management exactly. A Cross-reference agent links what one document says to what another allows — a revolver that springs ahead of the notes, an IP blocker that misses a non-guarantor subsidiary — and a Citation Checker recomputes every figure and removes any sentence it cannot cite. The answer is pinned, sentence by sentence, to the capital structure; the analyst decides low-confidence sentences, pushes the answer into a credit memo, and a senior analyst approves it before the portfolio manager reads it.
Transcript
- AI Credit Analysis answers credit analysts' questions across filings, indentures, calls, and notes cited.
- Analysts cover up to 80 issuers, each with 300-page agreements.
- Nadia asks, Can Halvorsen put new debt ahead of our 2029 notes?
- The Planner splits it into four questions and waits.
- Agents read public side sources only and pin each sentence to its tranche.
- $350 million free and clear on page $142 plus $128 million of ratio debt.
- The cross-reference agent sees the revolver spring ahead of our notes.
- A Norwegian report puts the patents outside the guarantors, Nadia confirms.
- Into the credit memo, Nadia submits it, and Daniel, the senior analyst, checks the flag sentences and approves.
- Cited, signed, and logged, with an evidence pack for auditors.
- Result, $631 million could rank ahead of us, found in 96 seconds.
- Built on Botminds: agents, data, screens, and governance on one platform.
Transcribed from the film’s narration. Captions are on the player.