Underwrite · PCT-067 · Sourced card

Mark gaps across managers

A valuation or credit team wants, for every borrower held by more than one public BDC, the spread of marks across managers, so it can test its own mark against the outliers rather than the average.

BDCsDirect lendingPublic dataBuildable in-house

Today vs automated

TODAY

Open 7 BDC 10-QsAssociate · manual
Match names in ExcelAssociate · manual, re-keying
Compute gaps by handVP / counsel · manual
Email / meeting hand-offAssociate → VP / counsel

AUTOMATED

BDC 10-Qs for the quarter are filedTrigger: starts on its own
Normalize namesAutomated · Document parser
Compute each gapAutomated · LLM
Flag the outliersAI draft, human review · Model template
Whether a gap reflects different tranches, different information or a different viewApprove / override · VP / counsel

AutomatedAI draft, human reviewHuman onlyTools: Data room, Document parser, LLM, Model template

Verified example

Lender / BDC
Borrower
Filing

How one borrower links three public lenders and the filings that report it. Verified example: SEC BDC Data Set, period ending 2026-06-30, checked against each filed schedule (ARCC, OBDC, GBDC 10-Qs). Marks are fair value ÷ cost for the position shown in the data set; OBDC and GBDC also list other tranches of the same borrower, not drawn. The borrower is named as each filing reports it; inclusion is not a credit view. Sponsor and agent bank are not disclosed in the schedules, so they are not drawn. Not investment advice.

Before: scattered rows

Where it sits	Row text (as filed)	FV ÷ cost
ARCC 10-Q, Schedule of Investments	Hyland Software, Inc. \| First lien senior secured loan	100.8%
OBDC 10-Q, Schedule of Investments	Hyland Software, Inc \| First lien senior secured loan \| Non-Affiliated	97.0%
GBDC 10-Q, Schedule of Investments	Hyland Software, Inc. \| One stop 1 \| Non-Affiliated Issuer	96.3%

Three filings, name strings that don’t match exactly, two different tranche labels. Nothing in the rows says they are the same company.

Question the graph answers

Which lenders share this borrower, and who marks it lowest?

One hop from Hyland Software: ARCC, OBDC and GBDC. Lowest mark: GBDC at 96.3% of cost. Two hops: the filing behind each mark.

Estimated valueestimate

16–88 analyst hours a year
$2k–$22k of time, vs an assumed $10k–$40k tool cost

Net: −$38k to $12k a year net of tool cost; at the low end of the assumptions it does not cover the tool cost.

Manual6–12 h per quarterly reviewWith AI1–2 h per quarterly review, review included

Assumptions

6–12 analyst hours per quarterly review today; 1–2 hours with an agent, including review
4–8 quarterly reviews a year for one team
$150–250 per hour loaded cost
Tool or build cost $10,000–40,000 a year (in-house)

Estimate, based on the assumptions shown; your numbers will differ. How we estimate

What else it can do for the team

Flag your positions with the widest peer gaps
Support valuation committee challenges
Draft the research note
Track gaps closing over time
Feed into secondary pricing

How you get thereFour steps from prompting to an agentic workflow; most teams start at step 2.

1 · PromptingAn analyst asks a chat tool about one quarterly review.
Needs: an approved AI tool, a confidentiality policy.
Risk: unchecked answers, data pasted outside policy.

2 · WorkflowA fixed pipeline extracts the same fields into your template every time.
Needs: field definitions, a template, document parsing.
Risk: breaks silently when formats change.
Typical starting point

3 · AgentIt watches the inbox or data room, runs the checks and drafts the output.
Needs: inbox/doc-store access, system write-back, logging.
Risk: acting on a misread; needs review gates.
The automated lane above

4 · Agentic workflowSeveral agents work across the deal or book; people approve exceptions.
Needs: shared data model, permissions, audit trail, exception queue.
Risk: errors compound across steps; needs monitoring.

Solution optionsBuy (3 vendors listed), build or hybrid, with the trade-offs of each.

Buy
Octus, The LoanEdge, Canoe Intelligence
Fastest start; you adapt to the product's data model and pricing.

Build
An LLM API, a document parser, a vector or document store, your data warehouse, and a review queue.
Fits your definitions exactly; you own maintenance and controls.

Hybrid
Vendor extraction feeding your own checks, templates and systems.
Common middle path; two things to integrate and govern.

Listing is not endorsement.

What changes for the team6–12 h → 1–2 h per quarterly review (estimate); coverage goes from the sections someone had time to read to every clause and schedule checked against the term sheet.

ROLESAssociates move from reading and retyping to checking cited exceptions.

CYCLE TIME6–12 h → 1–2 h per quarterly review (estimate)

COVERAGEFrom the sections someone had time to read to every clause and schedule checked against the term sheet.

CONTROLSEvery output cites its source; approvals and overrides are logged.

0same-loan pairs ≥10 pts apart

1near misses

16held out as incomparable

7reconciled BDCs

Source: SEC BDC Data Sets, period ending 2026-06-30. Derived from SEC filings; may contain extraction errors. Not investment advice. See the pairs →

How it works

Full title: Same borrower, different managers: where Ares, Blackstone and Blue Owl books agree and where they don't

“On 208 shared borrowers, public BDC marks are usually within a point. 21 aren't.”

Scenario

Public: a valuation or credit team wants, for every borrower held by more than one public BDC, the spread of marks across managers, so it can test its own mark against the outliers rather than the average.

What the human still decides

Whether a gap reflects different tranches, different information or a different view. Aggregated issuer-level marks can mix tranches; the human checks the rows.

Why it matters

The SEC staff statement of 28 Sep 2026 asks funds to calibrate fair value to transactions and current market evidence. Another manager's public mark on the same loan is that evidence.

What the record shows

Checked 2026-10-16.

Across the 208 borrowers held by two or more of the 7 reconciled BDCs at 30 Jun 2026 (debt only, FSK unfunded commitments excluded, names normalized), the median gap between the highest and lowest mark is 0.9 points, and 109 are within 1 point. Dispersion is concentrated: 21 borrowers show gaps of 5 points or more and 10 of 10 or more. By manager pair, BCRED–OBDC share 49 borrowers with 5 gaps of 5+ points, ARCC–BCRED 57 with 2, ARCC–OBDC 37 with 3, and ARCC–GBDC 67 with 7. One borrower held by four of them ($623m combined) sits at 93.3% (OBDC) to 102.4% (GBDC) of cost.

Try it: prompt and data

For every borrower held by two or more of ARCC, BCRED, FSK, GBDC, HTGC, MAIN and OBDC at 30 Jun 2026, list each holder's FV/cost, the max–min gap and the filing rows. Sort by gap. Flag where the holders' rows are different tranches.