Screen deal · PCT-004 · Sourced card
Screen a CIM in minutes
Reads a CIM against the fund’s investment box and drafts a go/no-go screen with page cites and a question list.
Worked example: Leverage is 6.6x on lender-adjusted EBITDA ($250m ÷ $38.1m) against a 5.5x fund limit. No-go as proposed; it screens in at about $210m. See the walk-through ↓
Today vs automated
- Read 70 pagesAssociate · manual
- Note EBITDA and leverageAssociate · manual
- Check the fund box by handVP · manual
- Email / meeting hand-offAssociate → VP
- A CIM or teaser lands in the deal inboxTrigger: starts on its own
- Pull key figuresAutomated · Document store
- Score each criterionAutomated · LLM
- Draft go / no-goAI draft, human review · Pipeline CRM
- Whether to spend more timeApprove / override · VP
AutomatedAI draft, human reviewHuman onlyTools: Deal inbox, Document store, LLM, Pipeline CRM
Estimated valueestimate
750–4,500 analyst hours a year
$112k–$1.1m of time, vs an assumed $20k–$100k tool cost
Net: $12k to $1.1m a year net of tool cost.
Assumptions
- 4–8 analyst hours per CIM today; 0.5–1.5 hours with an agent, including review
- 300–600 CIMs 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
- Build a first-pass comps set from your past deals and public lender marks
- Pre-fill the IC screening memo with figures and page cites
- Flag missing data-room items and draft the first diligence question list
- Check the deal against the fund box: sector, EBITDA size, leverage and hold limits
- Log the deal to the pipeline CRM with tags and the decline reason if passed
Worked walk-throughKey catch: Leverage is 6.6x on lender-adjusted EBITDA ($250m ÷ $38.1m) against a 5.5x fund limit. No-go as proposed; it screens in at about $210m.
1. The input
SAMPLE: fictional company, illustrative figures. Not a real borrower, deal or lender group.
A CIM executive summary and the fund’s investment box. SAMPLE deal pack →
“Project Meridian is a leading outpatient physical therapy platform with 142 clinics across 6 states. FY2025 revenue of $286.4m; Management Adjusted EBITDA of $41.8m (14.6% margin), including $3.7m of run-rate contribution from 2025 de novo clinics and identified procurement savings. Payer mix: 78% commercial, 14% Medicare, 8% other. The Sponsor seeks a $250m unitranche (6.0x) to refinance existing debt and fund a $25m distribution.”
CIM, Executive Summary pp. 4–6 (SAMPLE).
| Fund box criterion (SAMPLE) | Limit |
|---|---|
| Sector | Healthcare services allowed; no reimbursement-only models |
| EBITDA size | $15m–$75m (lender-adjusted) |
| Total leverage | Up to 5.5x lender-adjusted EBITDA |
| Use of proceeds | Dividend recaps limited to 10% of the facility |
| Single-name hold | Up to $40m |
2. What the AI produces
| Criterion | What the CIM says | Result | Cite |
|---|---|---|---|
| Sector | Outpatient therapy, 78% commercial payers | Pass | p. 4 |
| EBITDA size | $41.8m management; $38.1m before run-rate items | Pass | p. 6 |
| Leverage | $250m ÷ $38.1m = 6.6x vs 5.5x limit (6.0x on management EBITDA) | Fail | p. 6 |
| Use of proceeds | $25m distribution = 10% of $250m, at the limit | Flag | p. 5 |
| Add-backs | $3.7m run-rate items are 9% of EBITDA; QoE not yet in the data room | Flag | p. 6 |
Draft screen: “No-go at 6.0x as proposed. Would screen in at about $210m (5.5x on $38.1m) subject to the QoE. Questions: QoE status; run-rate savings actions taken; payer contract renewals.”
3. What a person still checks
- Recompute EBITDA once the QoE lands; the CIM’s figure is the sponsor’s.
- Decide whether a counter at lower leverage is worth the team’s time.
- Check sector and concentration limits against the live portfolio, not just the box.
4. Where it breaks
- CIMs bury adjustments in footnotes and appendices; a summary-only read misses them.
- The fund box changes with portfolio concentration; a static box goes stale.
- A pass on the box is not a credit view.
How you get thereFour steps from prompting to an agentic workflow; most teams start at step 1.
Needs: an approved AI tool, a confidentiality policy.
Risk: unchecked answers, data pasted outside policy.
Typical starting point
Needs: field definitions, a template, document parsing.
Risk: breaks silently when formats change.
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 (5 vendors listed), build or hybrid, with the trade-offs of each.
Hebbia, V7 Go, Indus Valley Partners (IVP for Credit), Claira, Raven (RavenRisk AI)
Fastest start; you adapt to the product's data model and pricing.
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 team4–8 h → 0.5–1.5 h per CIM (estimate); coverage goes from a sample of deals read in full to every inbound deal screened the day it arrives.
How it works
Full title: Kill the deal in 10 minutes: the agent that screens every CIM against your fund's box
“This CIM fails our box on page 41. It took the agent 90 seconds to find it.”
Scenario
SAMPLE: The Project Meridian CIM (fictional outpatient therapy platform) arrives with a $250m unitranche ask. The fund box says total leverage up to 5.5x lender-adjusted EBITDA, dividend recaps capped at 10% of the facility, and a $40m hold. The associate has three other CIMs in the queue.
What the human still decides
Whether to spend more time. The agent never says yes, it only shows which boxes fail and why.
Why it matters
CIM review is the first step of every deal and a recurring time sink. PwC’s 2026 survey lists data-room summaries and document extraction among the strongest AI uses today.
What to verify
Checked 2026-10-15.
Most CIMs describe private companies, so the company's own numbers cannot be checked against a public filing. What can be checked while screening: the sponsor, from its Form ADV (assets under management, funds, disciplinary history) and its funds' Form D notices on EDGAR; and the CIM's own claims, listed for diligence rather than accepted. The checklist: adjusted EBITDA and every add-back (how much is run-rate or pro forma, and whether it survives the credit agreement's definition, see the add-back autopsy card); customer and supplier concentration (top-10 share and contract terms); pro forma leverage on the lender's EBITDA, not the sponsor's; quality-of-earnings scope and who prepared it; and anything the CIM leaves out, which goes on the question list.
Try it: prompt and data
Here are our fund criteria: {criteria}. Read the attached CIM. For each criterion output pass/fail/unclear, the figure, and the page number. Then list the 5 most important unanswered questions. Do not recommend a decision.Data
Demo uses clearly labeled SAMPLE (fictional) documents. Never paste confidential or non-public deal data into a consumer AI tool.
Who uses it, vendors and build
- Segments
- Direct lending
- Who uses it
- Originator, Associate
- Build
- Vendor + in-house
- In-house
- LLM with long-document input + a criteria template + 10–20 labeled past CIMs as an eval set. Estimate: 1–2 weeks for a small team.
- Last verified
- 2026-10-15
Vendors that claim this capability
- Claira: cim-screening
- Hebbia: cim-screening
- Indus Valley Partners (IVP for Credit): cim-screening
- Raven (RavenRisk AI): cim-screening
- V7 Go: cim-screening
Tags reflect each vendor's own public claims at the linked page (checked 2026-10-15). Not tested, ranked or endorsed.
Sources and related
- SEC EDGAR full-text search (Form D, fund filings)
- SEC Investment Adviser Public Disclosure (Form ADV)
- Viavi Solutions term loan credit agreement, EX-10.1 (filed 30 Oct 2025)
- Credit agreement scan demo (every row hand-read against the filing)
Part of the deal lifecycle: stage 1 of 7, Inbound screening and the fast 'no'.
Next step
Get new walk-throughs
Not investment advice. Names marked SAMPLE are fictional. Found an error? See the corrections log.