What Is a Credit Assessment Memo? Structure, Sections and a Worked Example

A credit assessment memo (CAM) is the single document a lender's approving authority actually reads and signs against. It converts raw borrower evidence — financial statements, bank statements, bureau reports, security documents — into a structured argument for or against a specific facility, at a specific amount, on specific terms. Everything else in the file supports it.

It is not a summary written after the decision. It is the decision record. A YuSight sample CAM runs 28 minutes end to end and carries 142 citations, every figure clickable back to the page of the source document it came from — which is the standard the rest of this page is written against.

Key facts

• A CAM typically runs 12 to 14 named sections, from borrower overview through to recommendation and approval hierarchy. Skipping any of them is the most common reason a memo is returned by credit committee.
• CBUAE Circular C 3/2024 STA (Credit Risk Management Standards), effective 30 November 2024, requires UAE licensed financial institutions to document leverage, debt service coverage, liquidity, net worth and operating cash flows in the credit file, and to keep files complete enough to establish the obligor's current financial condition (Article 5.12–5.13).
• In India, RBI/2018-19/87 dated 5 December 2018 mandates a minimum 60% loan component for borrowers with aggregate fund-based working capital limits of ₹1,500 million (₹150 crore) and above, effective 1 July 2019 — a structural constraint your facility section must reflect (RBI).
• Writing a mid-market commercial CAM by hand takes roughly 15–24 analyst hours across spreading, bank statement analysis, ratio work and drafting.
• YuSight's CAM Generation module produces a 28-minute draft CAM carrying 142 citations, with 100% of figures cited and one-click source verification.

What is a credit assessment memo, and what is it not?

A credit assessment memo is a lending document. It answers four questions in order: who is borrowing, what are they borrowing for, can they repay it, and what happens if they don't. Every section exists to answer one of those four.

The memo has an audience with authority. A relationship manager writes a proposal; a credit analyst writes the CAM; a sanctioning authority — branch head, zonal credit committee, management credit committee, board — signs it. The document therefore has to survive being read by someone who has never met the borrower and will not read the underlying documents.

What it is not: a credit analysis. Credit analysis is the work. The CAM is the artefact. You can do excellent analysis and produce a memo that gets rejected because the covenant table is missing or the deviation is buried on page 11 instead of flagged in the recommendation. You can also produce a beautifully formatted CAM on top of a spread with a mis-keyed depreciation figure, and the whole thing is worthless.

Why do searches for "credit memo" return accounting software?

This is worth addressing directly, because it derails a lot of people.

There are two unrelated documents sharing a name.

The accounting credit memo (properly, a credit note) is issued by a seller to a buyer to reduce an amount already invoiced — for returned goods, a pricing error, a short shipment, a volume rebate. It is a receivables adjustment. Corporate Finance Institute's "Credit Memorandum" page defines it exactly this way: "a document used to reduce the amount owed by a buyer." Search results for the bare term "credit memo" are dominated by accounts-receivable and invoicing software for the same reason.

The lending credit memo is issued by a lender's credit function to its own approving authority, recommending that a facility be granted, renewed, enhanced or declined. Nobody's invoice changes. It is a credit decision document.

They share three syllables and nothing else. If you are a credit analyst and a colleague sends you an accounting-software comparison, that is why. To avoid the collision entirely, use the fuller terms: credit assessment memo, credit appraisal memorandum, or credit approval memorandum.

What is a credit assessment memo called in India, the US and the UAE?

The document is close to universal. The vocabulary is not.

• India — Credit Assessment Memo or Credit Appraisal Memorandum, almost always abbreviated CAM. Some public sector banks call it the credit proposal note or simply the appraisal note. NBFCs frequently say "CAM" for retail and MSME files too, even where the memo is two pages.
• United States — credit memorandum, credit write-up, credit approval package, or in smaller community banks just "the write-up". "CAM" is not standard US usage and will not be understood.
• UAE — Credit Application or Credit Memorandum, framed around the CBUAE Credit Risk Management Standards. Many UAE banks run an India-influenced or UK-influenced house format depending on parentage, so "CAM" is understood in some institutions and not in others.

The structure below holds in all three. The section names shift; the questions they answer do not.

What are the standard sections of a credit assessment memo?

1. Borrower and facility overview

Legal name, constitution, incorporation date, registered and operating addresses, identifiers (CIN/PAN/GSTIN/Udyam in India; EIN and Secretary of State registration in the US; trade licence number, VAT TRN and Emirates ID of signatories in the UAE), group structure, and existing exposure with your institution and others.

Then the ask, in a box: facility type, amount, tenor, pricing, security, and whether it is fresh, renewal, enhancement or restructure. A committee member should know the entire proposition inside 30 seconds.

2. Purpose of the facility

The single most under-written section. "Working capital requirement" is not a purpose. "To fund a ₹27.30 crore working capital gap arising from an 85-day cash conversion cycle on projected FY27 sales of ₹108 crore" is a purpose. For term loans: the asset, the supplier, the quotation, the project cost, and the means-of-finance split.

End-use monitoring obligations flow from this section. If the purpose is vague, the covenant testing that follows is unenforceable.

3. Promoter and management background

Promoter names, shareholding percentages, background and years in the industry, other directorships and their financial health, and the personal net worth statement of anyone whose guarantee you are taking. Include succession — a 68-year-old promoter with no second line is a risk factor whether or not anyone wants to write it down.

Negative checks belong here: defaulter lists, wilful defaulter checks, litigation, adverse media, sanctions screening.

4. Industry and business analysis

What the borrower sells, to whom, at what margin, and what would stop them. Customer concentration with actual percentages. Supplier concentration. Order book or contracted revenue. Capacity and utilisation. Competitive position — not "the industry is growing" but who else bids for the same tenders and on what basis they win.

The OCC's Commercial Real Estate Lending handbook (Version 2.0, March 2022) frames the equivalent expectation for CRE as evaluating "borrower and guarantor creditworthiness and global financial condition, including assets (type, amount, and liquidity), global cash flow, direct and contingent liabilities" (OCC).

5. Financial analysis and spreads

Three years audited, plus provisional or stub-period figures, plus projections for the facility tenor. Standardised into a common template so the numbers are comparable to peers and to prior years — this is spreading, and it is where most CAM errors originate.

Auditor's name, audit qualifications, changes in accounting policy, related-party transactions, and contingent liabilities all belong here. A qualified audit opinion that appears nowhere in the memo is a documentation failure regardless of the credit outcome.

6. Ratio analysis

Computed, trended, and interpreted. A ratio table with no commentary is decoration. Minimum set: growth, EBITDA margin, PAT margin, interest coverage, DSCR, debt/EBITDA, TOL/TNW, current ratio, and the working capital cycle broken into inventory, receivable and payable days.

7. Banking conduct and repayment track record

Bank statement analysis across all operating accounts: average utilisation of existing limits, frequency and duration of overdrawing, cheque returns inward and outward, minimum and average balances, and cash deposit patterns against declared turnover.

Bureau conduct: DPD buckets across all live facilities, enquiry velocity, restructuring flags, and any settled or written-off accounts. In India, CRILC reporting for aggregate exposure of ₹5 crore and above. In the UAE, AECB plus Al Etihad returns. In the US, business credit files plus personal scores of guarantors.

The single highest-value cross-check in the entire memo: does declared turnover reconcile to credits in the bank statements, and to GST/VAT returns?

8. Security and collateral

Primary security, collateral security, and guarantees — each with description, ownership, valuation, valuer name and date, and the resulting cover ratio. Charge type and perfection status: hypothecation, mortgage (registered or equitable), pledge, UCC-1 filing, MOA registration with the Registrar of Companies.

Unperfected security is unsecured lending with extra paperwork. Say so if it applies.

9. Risk factors and mitigants

Written as pairs, always. A risk with no mitigant is an argument to decline; a mitigant with no named risk is filler. Three to seven genuine risks, ranked. If your top risk is "general economic conditions", you have not done the section.

10. Covenants proposed

Financial covenants with the exact formula, the threshold, the test frequency and the first test date. "Maintain adequate DSCR" is not a covenant. "DSCR ≥ 1.35x, tested annually on audited financials, computed as (PAT + depreciation + interest on term debt) ÷ (interest on term debt + scheduled term debt repayment), first test on FY27 audited accounts" is a covenant.

Non-financial covenants too: end-use certification, no additional debt without consent, promoter shareholding floor, no disposal of secured assets, insurance maintenance with bank clause.

11. Pricing and terms

Rate and benchmark linkage, spread, reset frequency, processing fee, commitment fee, prepayment terms, repayment schedule with moratorium, and the risk-adjusted return calculation your institution uses. Pricing should visibly connect to the internal rating assigned earlier in the memo. If a rating-4 borrower is priced like a rating-2 borrower, the memo must explain why.

12. Deviations sought

Every departure from board-approved policy, listed explicitly with the policy clause, the requested deviation, the justification, and the authority competent to approve it. Deviations found buried in narrative rather than tabulated is the fastest route to an audit finding.

13. Recommendation and approval hierarchy

An unambiguous recommendation — approve, approve with conditions, or decline — followed by the conditions precedent and conditions subsequent, and the signature chain with delegated authority limits. CBUAE's standards require board or senior management involvement in approving materially large facilities.

What does the financial analysis actually look like? A worked example

Illustrative example. Vellore Precision Components Private Limited is a hypothetical borrower constructed for this article. The figures are invented to demonstrate method. They are not a real company, a real sanction, or a benchmark.

The borrower. Tier-2 automotive component manufacturer, machined transmission housings, incorporated 2009, two promoter-directors. Single plant. Top three customers are 61% of revenue.

The ask. ₹18.00 crore total — ₹12.00 crore fund-based cash credit (renewal plus enhancement) and ₹6.00 crore term loan for a CNC machining line costing ₹8.00 crore, funded 75:25 with ₹2.00 crore promoter contribution. Term loan: 5 years door to door, 6-month moratorium.

Spreads (₹ crore)

FY2026 balance sheet detail: inventory ₹14.80 cr, receivables ₹19.60 cr, cash ₹1.20 cr, other current assets ₹2.10 cr. Payables ₹13.40 cr, working capital borrowings ₹11.40 cr, current portion of long-term debt ₹3.60 cr, other current liabilities ₹2.30 cr. Raw material consumed ₹58.60 cr.

Ratios, computed line by line (FY2026)

Current ratio Current assets = 14.80 + 19.60 + 1.20 + 2.10 = ₹37.70 cr Current liabilities = 13.40 + 11.40 + 3.60 + 2.30 = ₹30.70 cr 37.70 ÷ 30.70 = 1.23x

Interest coverage 10.85 ÷ 3.42 = 3.17x

Debt / EBITDA 31.80 ÷ 10.85 = 2.93x

Total debt / TNW 31.80 ÷ 24.60 = 1.29x Adjusted for promoter unsecured loans subordinated by written undertaking: (31.80 − 6.20) ÷ (24.60 + 6.20) = 25.60 ÷ 30.80 = 0.83x The memo must state which of these two numbers the covenant tests against. Both are defensible; only one can be enforced.

Working capital cycle Inventory days = (14.80 ÷ 58.60) × 365 = 0.2526 × 365 = 92 days Receivable days = (19.60 ÷ 94.30) × 365 = 0.2078 × 365 = 76 days Payable days = (13.40 ÷ 58.60) × 365 = 0.2287 × 365 = 83 days Cash conversion cycle = 92 + 76 − 83 = 85 days

DSCR — the number the committee will argue about

FY2027 is the first full year after drawdown. Term debt repayment that year is ₹4.40 crore (₹3.20 crore existing term loans plus ₹1.20 crore on the new facility after the moratorium).

Method A — Indian cash-profit convention

DSCR = (PAT + Depreciation + Interest on term debt) ÷ (Interest on term debt + Term debt principal)

Numerator = 3.97 + 3.55 + 2.28 = ₹9.80 cr Denominator = 2.28 + 4.40 = ₹6.68 cr DSCR = 9.80 ÷ 6.68 = 1.47x

Method B — EBITDA over total debt service (common US and UAE convention)

Numerator = EBITDA = ₹12.90 cr Denominator = total interest 4.05 + term principal 4.40 = ₹8.45 cr DSCR = 12.90 ÷ 8.45 = 1.53x

Method C — Method B less unfunded maintenance capex and distributions

Assume ₹0.90 crore maintenance capex not funded by the term loan, and a ₹0.30 crore dividend. Numerator = 12.90 − 0.90 − 0.30 = ₹11.70 cr DSCR = 11.70 ÷ 8.45 = 1.38x

Same borrower. Same year. 1.47x, 1.53x, 1.38x. Against a 1.35x covenant all three pass, but the headroom differs by 13 percentage points of coverage — and against a 1.40x covenant, Method C fails. This is why a CAM must state the DSCR formula in the covenant section rather than quoting a bare number. It is also the single most common source of disagreement between a relationship team and a credit team.

India-specific: working capital limit assessment

Projected FY2027 current assets ₹42.50 cr; non-bank current liabilities ₹15.20 cr. Working capital gap = 42.50 − 15.20 = ₹27.30 cr Minimum 25% margin from long-term sources = 27.30 × 0.25 = ₹6.83 cr Maximum permissible bank finance = 27.30 − 6.83 = ₹20.47 cr

Actual net working capital = 37.70 − (30.70 − 11.40) = 37.70 − 19.30 = ₹18.40 cr, comfortably above the ₹6.83 crore minimum margin. The ₹12.00 crore request sits well inside the assessed MPBF, which the memo should say out loud — an under-utilised assessment is a fact the committee should see, not a gap to hide.

How does a CAM change across markets and facility types?

Facility type changes the weighting, not the skeleton:

How long does a credit assessment memo take, and where does the time go?

Elapsed time and analyst time are different problems. Elapsed time is dominated by document chase — three to ten working days waiting on the borrower, and no software fixes that. Analyst time is the part you control.

The lopsidedness is the point. Roughly two-thirds of manual CAM effort is transcription and arithmetic — keying figures out of PDFs into a spread template, recomputing ratios, chasing which of four versions of the balance sheet is current. None of it is credit judgement. The judgement work — deciding whether 61% customer concentration is tolerable at 2.93x leverage, whether the promoter's second line is credible, whether a 1.38x DSCR under Method C is acceptable — is perhaps three of those hours, and it is the only part a committee is actually paying for.

Rework is the hidden cost. A memo returned by committee because a figure could not be traced to source costs the elapsed time of a full committee cycle. That is why citation density matters more than drafting speed: YuSight's sample CAM carries 142 citations across a 28-minute draft, with 100% of figures traceable to the source document and page in one click.

What separates an approval-ready CAM from one that gets sent back?

Six failure modes cover most returns:

1. An untraceable number. A figure appears in the ratio table that exists nowhere in the spreads. Fatal, and instantly visible to an examiner.
2. A recommendation that does not follow from the analysis. Risks listed on page 8 that the recommendation on page 12 never addresses.
3. A covenant without a formula. Unenforceable at the first test date, and nobody notices until then.
4. A deviation not flagged as a deviation. Turns a credit question into an audit finding.
5. Turnover that does not reconcile. Declared sales versus bank credits versus GST/VAT returns, unexplained.
6. Stale security valuation. A valuer report older than policy permits, cited without the date.

A quality rubric worth running scores each section on completeness, evidence (is every figure sourced), and reasoning (does the section reach a conclusion). Three levels, thirteen sections — enough to make reviewer feedback specific rather than "needs more detail".

Key takeaways

• A credit assessment memo is the decision record, not a summary of one. Twelve to fourteen sections, each answering who, why, can they repay, and what if they don't.
• The lending CAM and the accounting credit note share a name and nothing else. Use "credit assessment memo" or "credit appraisal memorandum" to stay unambiguous.
• State the DSCR formula, not just the DSCR. The same borrower produced 1.47x, 1.53x and 1.38x on three legitimate conventions.
• Risks and mitigants are written as pairs. Covenants carry formulas, thresholds, frequencies and first test dates. Deviations are tabulated, never buried.
• Two-thirds of manual CAM effort is transcription and arithmetic. That is the part worth automating; the judgement is not.

Every figure in a CAM should be one click from the page it came from. That is the standard YuSight's CAM Generation module is built to — a 28-minute draft, 142 citations, full version history, and an analyst who spends their hours on the credit call instead of the keying.

See your first CAM in 30 minutes — book a live demo.

FAQ

What should a credit memo include?

Borrower and facility overview, purpose, promoter and management background, industry and business analysis, financial spreads, ratio analysis, banking conduct and repayment track record, security, risks and mitigants, covenants, pricing, deviations, and the recommendation with approval hierarchy. Thirteen sections is the working minimum for a commercial file.

What is the difference between a credit memo and credit analysis?

Credit analysis is the work you do; the credit memo is the document you hand to the approving authority. Good analysis can still produce a memo that gets returned, usually because a covenant has no formula or a deviation was never flagged.

What is the difference between a credit memo and a loan application?

The loan application comes from the borrower and states what they want. The credit memo comes from your credit function and states what you should give them, on what terms, and why — including everything the borrower would rather you did not write down.

What is a credit memo in commercial lending, and why do search results show accounting software?

In commercial lending it is the memorandum recommending a facility to a sanctioning authority. The search results are about the accounting credit note — a seller reducing a buyer's invoice — which is an entirely different document that happens to share a name.

What makes a credit memo approval-ready?

Every figure traceable to a source document, every risk paired with a named mitigant, every covenant carrying a formula and a first test date, every policy deviation tabulated, and a recommendation that visibly follows from the analysis above it.

How long does writing a commercial credit memo take manually?

Roughly 15 to 24 analyst hours to a first draft for a mid-market file, plus three to ten working days of elapsed time waiting on borrower documents. About two-thirds of the analyst hours go to transcription and arithmetic rather than judgement.

How do you standardise credit memos across analysts?

Fix the template, fix the ratio definitions, and fix the evidence rule — every number cites its source. Most inconsistency between analysts is not disagreement about credit, it is two people computing DSCR differently and neither writing down which formula they used.

What is a credit memo quality rubric?

A scoring sheet that rates each section on completeness, evidence and reasoning. It turns "this needs more detail" into "section 7 has no cheque-return data and section 10 has a covenant with no test date", which an analyst can actually act on.

Why does credit memo quality depend on spreading quality?

Every ratio, every covenant threshold and the entire recommendation sit on top of the spread. A mis-keyed depreciation line changes PAT, changes DSCR, and changes whether the borrower passes a covenant — and the memo will look perfectly well-argued the whole way down.

Can AI write a credit memo?

It can draft one — classify the documents, spread the financials, compute the ratios and produce the narrative. It should not make the credit call, and it should not produce a single figure that an analyst cannot click back to the source page it came from.