Benefit Street Partners answers questions faster with Codex

Anant Kumar builds tools to compare loan valuations and analyze credit agreements, using Codex for code review and faster follow-ups.

First BDC mark tracker version; ~10–12 total development hours, including refinements.

First-pass credit agreement analysis, followed by 1+ hour of human review.

Share

Building his own dashboardsBuilding his own dashboards

Building a credit agreement analyzer

Results at a glance

What’s next

Building his own dashboards

Building a credit agreement analyzer

Results at a glance

What’s next

At Benefit Street Partners, a credit-focused investment manager, answering an investment committee’s follow-up questions used to take days or weeks. Portfolio manager Anant Kumar says analysts can now turn a round of comments in a day with Codex, and often answer questions during the meeting itself. BSP is using these tools to add more review capacity and rigor to underwriting within the same timeframe, rather than to do more deals.

To create a dashboard he could share with colleagues, Kumar used to need someone with software expertise to build a website and pull in the data. Now he builds his own tools, using AI to draft code and review it for errors.

His tracker uses public SEC filings from business development companies (BDCs) to compare how firms value the same loan. The team can see how its marks compare with other holders’ valuations.

Kumar built the tracker and used Codex for code review. The first version took a day. After adding features and refining the interface over subsequent weeks, he estimates total development time at 10–12 hours total.

Kumar’s other tool analyzes credit agreements, which are often 300 pages or longer. Investors need to find clauses that could shift value away from lenders, a task that previously meant asking outside counsel to work through the agreement page by page. Kumar brought together memos from leading law firms on clauses and risks to watch for. He describes the analyzer as a “second brain within our organization.”

An uploaded credit agreement takes about 20 minutes to process into a first-pass analysis, followed by an hour or more of human review. BSP has refactored the tool with GPT‑6 Astra, and it is now live in the company-hosted portal.

“The AI is very helpful in collating data, ingesting it, synthesizing analysis, but the ultimate investment decision needs to be made by a human,” he says.

Underwriting is typically a multi-round process. A round consists of further diligence and a revised memo responding to the investment committee’s questions. Deal timelines range from a few days to a few weeks, so the number of rounds varies. Kumar says, “We feel more confident in our decision making because we have more information at our disposal.”

Built the first BDC tracker in a day, using Codex for code review; about 10–12 hours total including later additions.

Produces a first-pass analysis of a credit agreement in about 20 minutes, followed by an hour or more of human review.

Answers committee follow-up questions faster with Codex, often during the meeting, according to Kumar.

BSP is working to connect information across deals. Continuous underwriting is a longer-term aspiration: the team could monitor loans as news happens and examine the effects across its portfolio, rather than re-underwriting only after an event.

“Codex has unlocked my day, because I have many more hours in the day to do more high-level work and less routine work,” says Kumar.