Credit Policy & Precedent Search: the rule in force, how the bank has used it, and the test worked on your deal
Junior analysts get cited answers without waiting on senior staff.
See one case, screen by screen ↓A search solution for credit analysts that answers from the bank's own credit policy, underwriting guides, past approved and declined memos and senior reviewers' comments. Each answer gives the rule in force, how the bank has used it and how those loans performed, and the policy test worked on the analyst's own deal — with every sentence opening at its source.
How to choose AI solutions for lending →
Updated
A commercial credit analyst writing a memo asks the questions every new analyst has asked: does this count toward leverage, what did we do last time, who can approve the exception, what will the reviewer ask? The answers sit in a 212-page credit policy, nine underwriting guides, thousands of past approved and declined memos and the comments senior credit officers left on them.
Large banks closed most of their formal credit training programmes years ago, and the lenders those programmes trained are retiring. So the questions go to the few senior people left — who are also the review bottleneck on every deal. An analyst who misreads a limit either kills a good deal or sends an exception to committee unlabelled, and the lesson the bank already learned sits in a memo nobody can find.
Estimated split, in hours, for one policy or precedent question in a commercial credit memo — by hand and with the solution.
Six specialist agents answer from the policy, past memos and reviewers' comments; the analyst writes and a senior approves exceptions.
Monday 12 October. Priya Raman, seven months in role, is writing two memos: Kestrel Precision Machining wants to buy the plant it leases, and her own Hollis & Crane acquisition goes to credit committee on Thursday. Here is her morning, screen by screen, in the working solution.
Relationship manager Gina Torres asks: “Can we go to 80% LTV on owner-occupied industrial?” The answer opens worked on Kestrel Precision Machining. The rule in force is §6.4.2, quoted word for word: the loan may not exceed 75% of the lower of the purchase price or the appraised value. Above 75%, to 80%, is a policy exception under §9.2.3; Renata Silva, senior credit officer for real estate, may approve it; never above 85%, the supervisory limit.
“Answered in 12 s · 6 agents · 11 sources · every sentence cited”
Every past decision sits on one scale against the policy line: 27 decisions on owner-occupied industrial since 2019 — 19 within 75%, 6 by exception, 2 declined above 80%. Three precedents follow, each with what the bank decided and how the loan has done since: Brightwater Machining and Haldane Fabrication at 80% by exception, both performing and never past due; Corwin Cabinetry refused the exception at 1.18x coverage and approved at 75%.
Renata Silva, on Brightwater: “Five points on owner-occupied is fine when the operating company carries the debt at 1.4x or better. Below that, no.”
On Kestrel's own numbers — $3,400,000 price, $2,720,000 requested, a $2,550,000 policy maximum — the exception is $170,000, five points, within §9.2.3. Global coverage is 1.45x. Priya drafts the exception paragraph: every clause carries its citation, and she can edit it before it goes into section 3 of the memo. One more click sends it to Renata Silva with the rule, the three precedents and the numbers.
“Added to section 3 of the Kestrel memo with 6 citations · exception labelled on page one”
Hollis & Crane Packaging is buying Pell Corrugated. Policy §4.3.4 lets synergies count only when documented by a quality-of-earnings report or signed contracts, and never above 10% of pro forma EBITDA. Of the $580,000 in the borrower's model, $423,000 is credited: leverage is 3.12x, inside the 3.25x grade 5 limit. With no synergies it is 3.43x — and that figure decides whether it is an exception.
Ardent Valve & Fitting was approved by exception at 3.40x with no synergies and a step-down covenant to 3.00x by year two — paid as agreed. Bramwell Paperboard was approved by credit committee on planned synergies with no quality-of-earnings report, and downgraded to grade 7 in March 2024 when only $80,000 of $350,000 arrived. Tanager Tool & Die was declined at 3.90x, with its largest customer at 41% of sales.
Walt Ostrander, on Ardent Valve: “Show me leverage with zero synergies. If it’s over policy at zero, call it an exception and say why we live with it.”
The full answer gives the short answer, the deal worked line by line, what it means under §9.2 and how the bank decided before. Beside it, each agent's step is shown with its time: the scoper read her draft of 9 October, the policy finder found §4.3.1 and §4.3.4, the precedent finder ranked 41 acquisition loans to manufacturers, and the checker confirmed 11 sentences against their passages and removed none.
Every citation opens at its passage, and the policy itself is one click away: Credit Policy v14, §4.3 Leverage, pages 58–62, current version, with the rules the answer rests on highlighted — the 3.25x grade 5 limit and the §4.3.4 sentence that synergies count only when documented, and in no case above 10% of pro forma EBITDA. Beside it: 64 questions from 9 analysts on this section in 90 days, and what changed in v14.
Precedents ranks past memos against Hollis & Crane: 41 acquisition loans to manufacturers since 2019, 29 approved and 12 declined. Nine were approved above the limit with no synergies — every one a §9.2 exception. Three were downgraded within two years, two of them credited synergies without a quality-of-earnings report. She compares her deal with Ardent Valve, Corliss Plastics and Tanager Tool & Die.
My deal shows the test line by line: pro forma EBITDA of $4,230,000, $423,000 of synergies credited, $14,500,000 of funded debt. Five tests pass and one is an exception: leverage with no synergies, over by 0.18x — within the 0.50x a senior credit officer may approve. Fixed-charge coverage is 1.31x without synergies. Alongside are the four questions Walt Ostrander is likely to ask, each with a cited draft answer.
The calculator never changes a deal figure and never assigns a risk grade.
Priya adds the cited leverage paragraph to section 6.2: 3.43x with no synergies, 3.12x capped, labelled a policy exception for senior credit officer approval, with the Ardent Valve step-down as the mitigant and Bramwell named as the difference. Six citations go in with it. Walt Ostrander decides on the exception; Priya did not have to wait on him to write it.
“Added to section 6.2 with 6 citations · exception labelled on page one”
The same two deals, seen by the five people around them — what their week looked like, and what it looks like now.
Six specialist agents answer from the policy, past memos and reviewers' comments; the analyst writes and a senior approves exceptions.
Reads the question and the analyst's open deal — industry, purpose, proposed grade and numbers — so every search starts from the right facts.
Finds the governing sections of the current credit policy and underwriting guides, and quotes the passage with section and page.
Ranks past approved and declined memos by purpose, industry, size and structure, and adds what happened after from annual reviews and grade changes.
Finds what senior credit officers wrote about the same point on similar memos.
Applies the policy tests to the analyst's own numbers — leverage with and without synergies, coverage, concentration and equity — and shows every line of the working with its source.
Checks every sentence of the answer against the passage it cites, removes anything unsupported, and logs a policy gap when three questions on one point find no policy text.
Writes the memo; a senior credit officer approves exceptions. The agents propose; a named person decides.
Analysts can ask in plain words, or tell it what to write. Every answer is cited to the policy section, memo or reviewer comment it rests on.
Can the acquisition synergies count toward leverage?
Partly. Synergies count only when documented, capped at 10% of pro forma EBITDA — $423,000 of the $580,000 on Hollis & Crane. That gives 3.12x, inside the 3.25x limit; with no synergies it is 3.43x, which policy treats as an exception.
Can we go to 80% LTV on owner-occupied industrial?
Not within policy: owner-occupied industrial stops at 75%. 80% is a five-point exception Renata Silva may approve. The bank has done it six times since 2019, always at 78–80% with coverage above 1.40x — all six are performing.
What will Walt ask about this memo?
From his comments on the four closest memos, expect: leverage with zero synergies, and why we live with it; why this is not Bramwell again; the contract end date for any customer above 20%; and coverage without the synergies. Each has a cited draft answer on your deal.
Add the leverage paragraph to my memo, with citations
Added to section 6.2: 3.43x with no synergies, 3.12x capped, labelled a policy exception for Walt Ostrander under §9.2, with the Ardent Valve step-down as the mitigant. All four of Walt’s likely questions are now answered in the memo.
13 screens from the working solution, on its sample data. Pick one to see it large.
What the answers may draw on, the rules every answer follows, and who answers what the library cannot — all settings.
| Setting | Default | Choose from |
|---|---|---|
| Sources searched | Credit Policy v14, 9 underwriting guides, approved and declined memos, reviewer comments, industry ratio guides | Each source on or off · indexed nightly at 02:00 |
| Relationship managers’ call notes | Not searched | On · Off |
| Every sentence carries a source | On · locked | Locked by the Chief Credit Officer |
| Cite only the current policy as current | On · locked | Older versions shown with version and dates |
| Never say a deal or an exception is approved | On · locked | Answers name who may approve, under §9.2 |
| Log a policy gap after 3 unanswered questions on one point | On · sent to Elena Marsh | On · Off |
| Leverage, acquisitions, covenants go to | Walt Ostrander | Walt Ostrander · Renata Silva · Elena Marsh |
| Real estate and construction go to | Renata Silva | Walt Ostrander · Renata Silva · Elena Marsh |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample credit data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (MIT Sloan, Itaú credit memo case — 2–4 hours of manual work per memo). Background: Datos Insights and Bank Director on hiring credit analysts and the end of formal credit training; OCC Comptroller's Handbook on policy-exception reports. People, companies and deals named on this page are fictional — characters and sample data in the working solution.
A search solution for credit analysts that answers from the bank's own credit policy, underwriting guides, past approved and declined memos and senior reviewers' comments. Each answer gives the rule in force, how the bank has used it and how those loans performed, and the policy test worked on the analyst's own deal — with every sentence opening at its source.
A policy search tells you what the rule says. This also shows how the bank has applied it — every past decision on the same question, which exceptions it granted, what the reviewer said and whether the loan held up — and works the test on the numbers in your memo draft.
No. Answers name who may approve under the exception section of the policy — a senior credit officer up to 0.50x over a leverage limit, credit committee beyond. The agents never approve an exception, assign a risk grade or say a reviewer agreed to something they did not.
It says there is no rule to quote, and writes no rule of its own. It shows what the memos did — for earn-outs, two memos that treated them differently — and the analyst can send the question to the named senior for that topic. After three unanswered questions on one point, a policy gap goes to the policy owner.
Not by default. Declined memos are visible to credit staff only, relationship managers see approved memos, and relationship managers' call notes are not searched. The precedent finder never shows a memo the user may not open.
Every question, its answer and its sources go into a training log by analyst and topic. The head of credit analysis sees how many questions the library answered, what went to a senior, which topics each analyst asks about most and how many memos reviewers sent back.
Yes. The analyst writes and signs the memo; agents draft cited paragraphs the analyst can edit before they go in. Exceptions are approved by whoever the policy names, and routing, sources and answer rules are settings — with the citation rule locked by the Chief Credit Officer.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your credit policy, underwriting guides and a sample of past memos and reviewer comments — and it goes live once every quality gate has passed. We will show it to you on your own policy first.
Every related entity spread in hours, each number linked to its page.
1-min demoCredit Memo AutomationA cited first draft the same day, not after days of pasting.
1-min demoCommercial Loan OriginationFiles ready for underwriting in days, not weeks of chasing documents.
1-min demoWe’ll run Credit Policy & Precedent Search on your own policy and a sample of your past memos.