Report · PCT-030 · Published note
Software exposure by lender
An LP wants to know how much of each BDC's book is software and how those marks are moving.
Today vs automated
- Open 10-Q schedulesAnalyst · manual
- Hand-map industriesAnalyst · manual
- Tie out totalsHead of research / IR · manual
- Email / meeting hand-offAnalyst → Head of research / IR
- Quarter-end, or a new filing is publishedTrigger: starts on its own
- Tag filer industry headingsAutomated · Data store
- Tie to industry tablesAutomated · LLM
- Chart share of bookAI draft, human review · Reporting template
- Whether a lender’s software share reflects its own labels or real exposureApprove / override · Head of research / IR
AutomatedAI draft, human reviewHuman onlyTools: Filing feeds, Data store, LLM, Reporting template
Estimated valueestimate
16–72 analyst hours a year
$2k–$18k of time, vs an assumed $20k–$100k tool cost
Net: −$98k to −$2k a year net of tool cost; on these assumptions the time saved does not cover the tool cost.
Assumptions
- 6–10 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 $20,000–100,000 a year (both)
Estimate, based on the assumptions shown; your numbers will differ. How we estimate
What else it can do for the team
- Tag your own book with the same method
- List software names with weak marks
- Track concentration trend
- Feed exposure into LP reporting
- Compare against sector limits
How you get thereFour steps from prompting to an agentic workflow; most teams start at step 2.
Needs: an approved AI tool, a confidentiality policy.
Risk: unchecked answers, data pasted outside policy.
Needs: field definitions, a template, document parsing.
Risk: breaks silently when formats change.
Typical starting point
Needs: inbox/doc-store access, system write-back, logging.
Risk: acting on a misread; needs review gates.
The automated lane above
Needs: shared data model, permissions, audit trail, exception queue.
Risk: errors compound across steps; needs monitoring.
Solution optionsBuy (2 vendors listed), build or hybrid, with the trade-offs of each.
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.
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 team6–10 h → 1–2 h per quarterly review (estimate); coverage goes from the names someone had time to pull to the full set, refreshed when sources change.
How it works
Full title: The SaaSpocalypse in the marks: which lenders hold the software loans AI might eat
“Software is about two-thirds of private credit's unrealised losses since Q4 2025. Here's who holds it.”
Scenario
Public: an LP wants to know how much of each BDC's book is software and how those marks are moving.
What the human still decides
Labels are each filer's own taxonomy (ARCC cites GICS 'Software and Services'; FSK uses 'Software & Services' without citing GICS; BCRED and GBDC use a narrower 'Software'), so shares are not like-for-like. MAIN does not disclose industry by holding. Facility caps use each facility's own definitions, not SOI headings. Spot-check rows against the filed HTML before quoting any share or mark.
Why it matters
BlackRock attributes about two-thirds of private credit's unrealised losses since Q4 2025 to software, with first-lien software marks around 97% of cost in Q2. Facility lenders are starting to cap software collateral (First Eagle, Oct 2026).
Before and after
Before
- Open each BDC 10-Q Schedule of Investments and search for software, SaaS, and related industry tags
- Copy matching rows into a spreadsheet and reconcile names that differ by filer
- Sum cost and fair value, compute software share of each BDC's book, and list names marked below 90% of cost
After (agent first draft)
- Agent loads the SEC BDC SOI table and each filer's 10-Q HTML Schedule of Investments for the quarter
- Tags every holding with its filer's own SOI industry heading (or FSK's Industry column), never from words in company names
- Checks each filer's tagged industry subtotals against its own industry composition table (within 0.5%)
- Computes per-BDC software cost, FV, share of book and names marked under 90% of cost, narrow and broad, and lists holdings linked to the filing
Agent pattern: download → tag from SOI industry headings → tie to filer industry table → aggregate by BDC → rank marks → cite filing → human spot-check
Tools
- Python or any agent with a shell
- SEC BDC Data Sets
- Static JSON on this site
Try it: prompt and data
From each filer's {quarter} Schedule of Investments, tag holdings with the filer's own industry heading or Industry column (no name matching). Tie tagged subtotals to the filer's industry table within 0.5%. Compute narrow and broad software share of FV per BDC and list software names below 90% of cost.Data
Public inputs only (SEC/EDGAR filings, published methodologies, public signals). Every figure is cited.
Who uses it, vendors and build
- Segments
- Research & ratings, BDCs, Direct lending
- Who uses it
- Credit analyst, LP / credit research
- Build
- Vendor + in-house
- In-house
- SOI pipeline + classifier prompt. Estimate: days.
- Last verified
- 2026-10-15
Vendors that claim this capability
Tags reflect each vendor's own public claims at the linked page (checked 2026-10-15). Not tested, ranked or endorsed.
Sources and related
Next step
Get new walk-throughs
Not investment advice. Names marked SAMPLE are fictional. Found an error? See the corrections log.