Direct Lending Credit Analysis: A 7-Step Workflow Using Claude AI

Direct Lending Credit Analysis: The Complete Workflow

A direct lending analyst covering a new middle-market deal is looking at a $25M to $250M EBITDA company, PE-sponsored, typically structured as a first lien term loan or unitranche at 4.0x to 5.5x leverage. The deal might close in 45 to 60 days from mandate. In that window, the analyst has to complete a full credit analysis — business quality assessment, financial statement normalization, debt capacity modeling, downside scenario analysis, credit structure review, sponsor evaluation, and a 25 to 35 page investment committee memo — while also working 5 to 10 other live deals in various stages.

The bottleneck is not the analytical judgment. It is the time required to extract, organize, and translate raw deal information — CIM, management presentations, 3 to 5 years of financials, quality of earnings report — into the structured format an IC memo requires. Claude handles the extraction and drafting so the analyst can spend more time on the judgment calls that actually determine whether the credit succeeds.

This guide follows a hypothetical deal throughout: the acquisition of Nexus Software, a $80M LTM EBITDA B2B SaaS business serving mid-market professional services firms, being acquired by Northgate Capital Partners in a sponsor-to-sponsor buyout. Deal terms: $800M enterprise value, $360M total debt (all unitranche), SOFR+550, 5% OID, 6-year term with 1% annual amortization, 5-year hold period. That is 4.5x leverage at close on a business with 74% gross margins, 92% net revenue retention, and $72M of ARR.

For related workflows, see our guides on credit agreement analysis, covenant compliance monitoring, and credit analysis with AI.

Step 1: Business Quality Assessment

Business quality is the first screen in any direct lending credit decision, and it determines what leverage the business can support and at what pricing. The key dimensions are: (1) competitive moat — does the company have pricing power, switching costs, or a captive customer base, or is it one of several interchangeable vendors?; (2) revenue visibility — what percentage of revenue is recurring, under contract, or highly predictable vs. project-based or transactional?; (3) customer concentration — is any customer above 10% of revenue, and what happens to EBITDA if that customer churns?; (4) management depth — is the business dependent on one or two key people or does it have a bench?; and (5) industry dynamics — is the industry growing, stable, or structurally declining, and are there regulatory or technology risks on the horizon?

For Nexus Software, the business quality case is strong: 92% NRR indicates low churn and expansionary revenue, the top 10 customers are 31% of ARR (no single customer above 7%), the product is embedded in client workflows with multi-year contracts, and the market (AI-augmented workflow software for professional services) is growing at 18% annually. The risk factors are management depth (CFO joined 14 months ago, prior CFO still on the board as an advisor), customer concentration in legal and accounting verticals, and the competitive threat from point solutions offered by larger platform vendors.

• "I am a credit analyst at a direct lending fund evaluating a new deal. Below is an excerpt from the CIM for Nexus Software, a B2B SaaS company serving mid-market professional services firms. Key facts: $80M LTM EBITDA, $110M ARR, 92% NRR, 74% gross margin, top 10 customers = 31% of ARR, no single customer above 7%, primarily 3-year contracts auto-renewing, management: CEO (10-year tenure), CFO (14 months), CTO (6-year tenure), 2 VP Sales (6 and 4 year tenure). Market: AI-enhanced document automation and workflow for legal, accounting, and consulting firms. 3 primary competitors: one large platform vendor with an adjacent product, one PE-backed pure-play (smaller), one open-source alternative used by price-sensitive buyers. From a credit perspective, assess: (1) revenue quality and visibility score (1-5 scale with explanation); (2) customer concentration risk and EBITDA impact if top 3 customers churn; (3) management depth and key-person risk; (4) competitive moat durability — is this a product with real switching costs or is it replaceable in a 12-month migration?; (5) top 3 credit risks from business quality that will be the focus of our underwriting diligence. Be specific — do not give generic SaaS risks."

Step 2: Financial Statement Normalization

The EBITDA on the cover page of a sponsor's CIM is almost never the number a direct lender will underwrite to. Sponsors present "Management Adjusted EBITDA" — GAAP EBITDA plus a list of add-backs — and the analytical work for the credit analyst is to assess each add-back for defensibility. The EBITDA bridge from GAAP to credit agreement EBITDA is the single most consequential calculation in the credit analysis, because it directly determines the leverage ratio and the covenant headroom at close.

For Nexus Software, Northgate Capital's CIM shows LTM GAAP EBITDA of $64M and Management Adjusted EBITDA of $80M — a $16M bridge. The add-backs include: sponsor management fee elimination post-close ($2.5M), non-recurring customer implementation costs for a large enterprise migration ($3.8M), stock compensation expense ($4.2M), one-time legal and compliance costs ($1.8M), and run-rate cost savings from a planned IT infrastructure migration ($3.7M). A conservative lender will not accept all $16M — particularly the run-rate IT savings, which have not been achieved and represent a forward-looking estimate. Getting the right EBITDA number changes leverage from 4.5x to something closer to 5.2x if the savings are haircut significantly, which is the difference between a deal that fits the fund's mandate and one that does not.

• "EBITDA normalization analysis for Nexus Software. The sponsor (Northgate Capital) claims Management Adjusted EBITDA of $80M for the LTM period ending Q1 2026. GAAP EBITDA is $64M. The sponsor's EBITDA bridge claims the following add-backs: (1) Sponsor management fee elimination: $2,500,000 — Northgate charges a $2.5M annual management fee that will be eliminated at close; (2) Non-recurring customer migration costs: $3,800,000 — one-time professional services and internal labor for migrating a 600-seat enterprise customer; (3) Stock-based compensation: $4,200,000 — non-cash; (4) One-time legal and SEC inquiry response costs: $1,800,000 — related to a routine SEC comment letter that was resolved; (5) Run-rate IT infrastructure savings: $3,700,000 — expected savings from migrating from co-located servers to AWS, migration is 40% complete as of Q1 2026. For each add-back: (A) assess whether a conservative direct lender would accept, haircut, or reject it, with reasoning; (B) apply your proposed haircut and state the conservative credit analyst's adjusted EBITDA; (C) calculate what leverage looks like at your adjusted EBITDA vs. the sponsor's $80M claim, given $360M of total debt. My fund's maximum underwriting leverage for a SaaS business is 5.5x — does this deal fit at your adjusted EBITDA?"

Step 3: Credit Structure Analysis

With normalized EBITDA established, the next step is analyzing whether the proposed credit structure is appropriate for the business. This means evaluating: the LTV ratio (total debt as a percentage of enterprise value), the debt capacity of the business at various leverage multiples, whether the unitranche or first/second lien structure is the right fit, and whether any PIK or cash pay toggle is appropriate given the cash flow profile.

For Nexus Software, the proposed structure is 4.5x leverage (at sponsor's EBITDA) on an $800M enterprise value — a 45% LTV, which is conservative by software standards. The all-first-lien unitranche at SOFR+550 is appropriate for this risk profile; a two-tranche structure would save the borrower 25-50bps on the first lien but add complexity and a second lien at SOFR+850+, making the blended cost similar or worse. The 5% OID brings all-in day-one economics to approximately SOFR+660 on a 5-year effective amortization basis — high for a quality SaaS business but reflective of current market conditions.

• "Credit structure analysis for Nexus Software. Proposed deal: $800M enterprise value (12x LTM Management EBITDA of $66M, my normalized figure), $360M unitranche at SOFR+550, 5% OID, 1% annual amortization, 6-year term, $40M delayed draw term loan available for add-on acquisitions. SOFR current rate: 4.50%. Analyze: (1) LTV ratio at close on my normalized EBITDA vs. sponsor's claimed EBITDA — at what EBITDA does LTV reach 50% (our maximum threshold for a software business)?; (2) all-in effective yield including OID amortized over 5 years, assuming SOFR remains at 4.50%; (3) comparison of unitranche vs. a first/second lien structure — model a 60/40 split ($216M first lien at SOFR+425, $144M second lien at SOFR+825) and calculate the blended cost vs. the unitranche; (4) annual cash interest burden at close on the full $360M facility; (5) interest coverage ratio (EBITDA / cash interest) at my normalized EBITDA; (6) minimum EBITDA needed to maintain 2.0x interest coverage — express as a percentage decline from current EBITDA. This comparison will inform our recommendation on structure."

Step 4: Cash Flow Modeling and Debt Service Analysis

The fundamental question in direct lending credit analysis is whether the business can service its debt in a downside scenario. Cash flow modeling for a credit investor is different from an equity investor's model: the focus is on the floor of cash generation, not the ceiling. Equity is modeled for upside optionality; debt is modeled for whether the company runs out of money in a stress scenario.

For Nexus Software, the key cash flow drivers are: ARR growth and churn assumptions (which drive revenue), gross margin stability (74% at close — can it sustain or does it compress with enterprise mix shift?), R&D investment intensity (SaaS companies typically run at 15-20% of revenue in R&D, which is a semi-fixed cost), and the capex profile (primarily capitalized development costs, minimal maintenance capex for a software business). In the base case the company can grow into its leverage quickly; the question is whether in a downside scenario — ARR growth stalls, a major competitor captures market share, or an integration issue causes higher-than-expected churn — the business can still service $19.8M of annual cash interest and cover the 1% annual amortization ($3.6M).

• "Cash flow and debt service model for Nexus Software. Starting point: LTM Revenue $148M, LTM EBITDA (my normalized) $66M, EBITDA margin 44.6%. Annual cash interest at SOFR+550 on $360M: $19.8M (assuming 4.50% SOFR). Annual amortization: $3.6M (1% of $360M). Capitalized development costs (treated as capex): $12M per year. Cash taxes: assume 25% tax rate on income after interest deduction. Model three 5-year scenarios: (A) Base case: ARR grows 12%/year, EBITDA margin stable at 44%; (B) Bull case: ARR grows 20%/year, EBITDA margin expands to 47% by year 3 as scale benefits flow through; (C) Bear case: ARR growth slows to 4%/year (churn increases to 12% from 8%, new logo wins slow), EBITDA margin compresses to 39% due to increased support and sales costs. For each scenario, show annually: (1) Revenue, EBITDA, EBITDA margin; (2) Less: Cash interest, Amortization, Capex (development costs), Cash taxes; (3) Free Cash Flow after debt service; (4) Cumulative deleveraging — total debt balance and leverage ratio at each year-end; (5) Whether FCF is sufficient to cover debt service in all years. Flag the year in the bear case when debt service coverage falls below 1.0x, if applicable."

Step 5: Sponsor Assessment

In PE-backed direct lending, the sponsor is not just the equity owner — they are a key risk mitigant and, if needed, the source of rescue capital. A sponsor with $3B of dry powder backing a $440M equity check can inject $30-50M of additional equity to buy time if the portfolio company runs into difficulty. A sponsor at the end of its fund lifecycle with little remaining dry powder may not be in a position to support the company through a covenant breach or liquidity crunch. The sponsor assessment covers four dimensions: track record with similar businesses, exit thesis credibility, financial capacity to support the company, and the relationship quality between sponsor and management.

Northgate Capital Partners, for Nexus Software, is a $6B AUM lower-middle-market software specialist on its Fund IV ($2.1B, vintage 2023). They have closed 12 software buyouts across Funds II and III; 4 have been realized at a median 3.2x MOIC and 28% IRR. The exit thesis for Nexus is a strategic sale to a large HR/ERP platform vendor within 4-5 years — plausible given recent M&A activity in the AI-enhanced workflow space. Fund IV is 40% deployed, meaning Northgate has approximately $1.3B of dry powder remaining across the fund's investment period.

• "Sponsor assessment for the Nexus Software credit. Sponsor: Northgate Capital Partners, Fund IV ($2.1B, 2023 vintage, 40% deployed, approximately $1.3B dry powder). Track record on prior software buyouts: 12 investments across Funds II and III, 4 realized at median 3.2x MOIC / 28% IRR, 8 unrealized with 2 on watch. Equity check in Nexus: $440M (55% of EV). Analyze from a credit perspective: (1) capacity and willingness to support — given $1.3B of dry powder and a $440M equity check, what is a realistic estimate of Northgate's ability to inject additional equity if Nexus underperforms?; (2) exit thesis assessment — is the thesis (strategic sale to ERP/HR platform) credible given the Nexus product, current M&A activity in the sector, and Northgate's typical hold period?; (3) what is the impact on credit risk if Northgate shifts from a 4-5 year exit to a longer hold — does this help or hurt the credit?; (4) any structural protection we should request given this is a software-focused sponsor (e.g., a technology-related MAC definition, IP ownership covenant, specific change of control protections)?"

Step 6: Downside Scenario and Covenant Breach Analysis

Every direct lending credit memo includes a downside scenario analysis showing how much EBITDA can fall before the borrower breaches its maintenance covenants, runs out of liquidity, or is unable to service debt. The covenant structure for Nexus Software includes a maximum leverage covenant of 5.25x at close stepping to 5.0x at year 2, 4.75x at year 3, and 4.50x at year 4 — set at approximately 17% headroom above closing leverage. There is also a minimum liquidity covenant of $15M (cash plus revolver availability) to prevent a sudden liquidity squeeze without a lender trigger.

The key question in the downside analysis is not "does the business survive?" — a stable SaaS business with 92% NRR is unlikely to go to zero. The question is "at what EBITDA level does the company breach its maintenance covenants, and how likely is that to occur?" For a fund lending on 5-year hold thesis, the scenario analysis should run across 3 to 4 scenarios at 1-year intervals to understand how the covenant headroom trajectory evolves under different business trajectories.

• "Downside scenario and covenant breach analysis for Nexus Software. Covenant structure: Maximum Total Net Leverage Ratio 5.25x at close (Q3 2026), stepping to 5.00x at Q3 2027, 4.75x at Q3 2028, 4.50x at Q3 2029. Closing leverage: 4.50x (my normalized EBITDA of $66M, total debt $360M, net debt $348M assuming $12M closing cash). Minimum Liquidity: $15M (unrestricted cash plus undrawn revolver). Total debt at close: $360M, amortizing at 1% per year. Run three downside scenarios from closing: (A) Mild stress: EBITDA grows at 5%/year (below sponsor's 18% growth case); (B) Moderate stress: EBITDA flat in year 1, then recovers at 3%/year from year 2; (C) Severe stress: EBITDA declines 20% in year 1 (ARR churn spike to 18%), then flat for 2 years, then recovers at 5%/year. For each scenario, show: (1) EBITDA, net debt (amortization applied), and leverage ratio at each year-end from close through year 5; (2) leverage ratio vs. applicable covenant at each step-down date; (3) in which scenario and which year does the company first breach its leverage covenant?; (4) the 'EBITDA cushion' at each step-down — how much EBITDA can decline from that point before breach?; (5) in the severe stress scenario, what is the minimum revolver draw needed to maintain $15M liquidity, and when does revolver capacity become the binding constraint? Present as a sensitivity table."

Step 7: Credit Memo Drafting — The IC Recommendation

The investment committee memo is the deliverable that synthesizes all of the prior analytical work into a recommendation. A well-structured credit memo for a direct lending deal follows a consistent format: executive summary with a clear recommendation at the top, business overview with competitive positioning, financial performance with normalized EBITDA bridge, credit structure with debt capacity analysis, risk factors (typically 4-6 specific risks with mitigants), sponsor assessment, and covenant package summary. The IC memo should be written so that a member of the investment committee who has not seen the deal can read the executive summary and have a clear view of the investment thesis and the key risks within two pages.

Claude's greatest value in credit memo drafting is generating the structured narrative sections from data you provide. The financial analysis section is particularly well-suited to Claude drafting — you give it the normalized EBITDA, leverage profile, coverage ratios, and downside analysis, and it produces a structured three-to-four paragraph analysis that meets IC memo standards. The risk section requires more analyst judgment — Claude can frame the risks, but the weighting and the mitigant assessment reflects your fund's credit philosophy.

• "Draft the Executive Summary and Credit Highlights sections of the IC credit memo for Nexus Software. Context: we are a direct lending fund recommending a $360M unitranche commitment at SOFR+550, 5% OID, 6-year term, 5.25x leverage covenant at close with step-downs to 4.50x by year 4. Key facts: $800M EV, $80M sponsor EBITDA ($66M our normalized), $148M revenue, 92% NRR, 74% gross margin, 12x EV/EBITDA, first lien only, sponsor is Northgate Capital (Fund IV, $1.3B dry powder, 40% deployed), exit thesis = strategic sale within 4-5 years to ERP/HR platform vendor. Our credit recommendation: APPROVE with conditions (financial covenant package as described, $15M minimum liquidity covenant, IP assignment and change of control protections). Write: (1) a 3-paragraph Executive Summary covering (a) the transaction and use of proceeds, (b) the credit investment thesis and why this is a compelling direct lending opportunity, (c) the key risks and why they are manageable; (2) a Credit Highlights table with 4-5 bullet points covering business quality, financial profile, structural protections, and sponsor; (3) a Key Risks table with 4 specific risks (no generic boilerplate — specific to Nexus Software), each with a 1-2 sentence mitigant. Write in a direct, IC-grade voice — no marketing language, facts and analysis only."

Putting the Workflow Together

The seven steps above represent 3 to 5 days of analytical work for an experienced analyst on a new deal. With Claude handling the extraction, normalization testing, and memo drafting, the timeline compresses to 1 to 2 days for the analyst's core time — the rest is judgment, diligence calls, and sponsor interaction that Claude cannot replace.

The most effective setup is a Claude Project for each active deal. Configure the project with: (1) the deal facts — company name, deal size, proposed structure, key financials — as the system prompt; (2) the CIM executive summary and management presentation pasted as reference documents; (3) the QoE report highlights if available. From that base, run each of the seven workflow prompts in sequence, iterating on each step's output before moving to the next.

For the credit agreement review that follows a mandate award, see our guide on credit agreement analysis with AI. For ongoing portfolio monitoring after close, see the covenant compliance monitoring workflow. The Private Equity and Commercial Banking skill categories include direct lending underwriting templates for each step of this workflow.

Frequently Asked Questions

What leverage multiples are typical in middle market direct lending today?

Middle market direct lending deals in 2025-2026 are generally closing at 4.0x to 5.0x total leverage for core middle market companies ($25M to $150M EBITDA). Software and high-quality SaaS businesses with strong NRR and recurring revenue can support 4.5x to 5.5x given higher enterprise value multiples and more predictable cash flows. Traditional industrials, distribution, and services businesses tend to underwrite at 3.5x to 4.5x. The direct lending market has applied modest leverage discipline since the 2021-2022 peak, and covenant packages in direct lending remain meaningfully tighter than broadly syndicated loans, which is a key structural distinction.

What is the difference between a unitranche and a first lien/second lien structure?

A unitranche combines first lien and second lien debt into a single instrument at a blended interest rate. Instead of a first lien at SOFR+425 and a second lien at SOFR+825, a unitranche at SOFR+575 gives the borrower one lender, one set of documents, and faster execution. Direct lenders prefer unitranche because it eliminates inter-creditor agreement complexity and gives the lender full control of the debt stack. Borrowers often prefer it for execution speed. The main alternative — a first lien / second lien split — allows borrowers to use the cheaper first lien more aggressively and bring in a separate second lien investor, which can lower the overall cost of capital for larger, higher-quality credits but adds inter-creditor complexity.

How much of a sponsor's claimed EBITDA add-backs should a direct lender accept?

Conservative direct lenders typically accept 60 to 80 percent of sponsor-claimed add-backs, applying haircuts based on defensibility and verifiability. Management fee eliminations at close are almost always accepted at 100 percent — they are contractual and certain. Non-recurring legal or consulting charges with clean documentation are generally accepted at 75 to 100 percent. Run-rate synergies and cost savings are the most scrutinized: lenders typically require 12 months of demonstrated operational history before giving credit for run-rate savings, and may accept only 50 percent of forward-looking operational improvements pending completion. Stock-based compensation is generally accepted as a non-cash add-back but is analyzed separately to understand dilution dynamics.

How long does a complete direct lending credit analysis take with AI assistance?

A conventional complete credit analysis for a middle-market direct lending deal takes 4 to 7 analyst-days from CIM receipt through IC memo submission — depending on deal complexity, financial history quality, and how much QoE support is available. With a Claude AI workflow covering EBITDA normalization, debt capacity analysis, cash flow modeling, downside scenario construction, and IC memo drafting, analysts consistently report compressing that timeline to 2 to 3 days. The time savings come primarily from the normalization and memo drafting steps, each of which typically takes half a day manually but can be completed in 30 to 60 minutes with well-constructed Claude prompts. Judgment-intensive steps — management calls, sponsor relationship assessment, legal review — are not accelerated by AI and remain the irreducible core of the analyst's time.