How AI Is Transforming MCA Underwriting: What Brokers Need to Know in 2026

The Underwriting Revolution Nobody Is Talking About

A few years ago, submitting a merchant cash advance file meant waiting. An underwriter at the funder would open three to six months of bank statements, manually calculate average daily balances, count negative days, look for NSFs, and come back to you — sometimes hours later, sometimes the next morning — with a decision or a list of stips.

That world is mostly gone now. The best-capitalized funders in the industry have deployed machine learning models that analyze thousands of data points and return a decision in seconds. Not minutes. Seconds.

For brokers, this is both an opportunity and a trap. If you understand how these systems work, you can package files better, set merchant expectations accurately, and stop submitting deals that were never going to close. If you don't understand them, you'll keep getting confused by fast declines on files that looked fine to you — and slow approvals on files that seemed complicated.

This guide breaks down what AI underwriting actually analyzes, why it prices deals the way it does, and how brokers can work with these systems instead of against them.

What Traditional MCA Underwriting Looked Like

The classic MCA underwriting checklist was simple by design. A human underwriter would look at:

• Three to six months of business bank statements
• Merchant processing statements (if the business took credit cards)
• A one-page application with time in business, industry, and ownership
• A soft or hard credit pull on the owner

The underwriter's job was pattern recognition — but human pattern recognition, which is slow, inconsistent, and subject to bias. Two underwriters at the same shop could look at identical bank statements and reach different conclusions about whether the deal was fundable.

AI doesn't have that problem. A well-trained model applies the same logic to every file, every time, at a speed no human can match.

What AI Underwriting Systems Actually Analyze

Modern MCA underwriting engines don't just tally deposits. They build a complete financial profile of the merchant from the transaction-level data in bank statements, then cross-reference that profile against industry benchmarks, macroeconomic signals, and the funder's own historical performance data.

Transaction-Level Cash Flow Analysis

When you submit bank statements to an AI-powered funder, the system isn't reading the PDF the way a human would. It's parsing every line — every deposit, every withdrawal, every fee — and extracting structured data from unstructured text. From that data, it calculates:

• Average daily balance (ADB): The single most important metric in MCA underwriting. Most AI models want to see an ADB that is at least 10 to 15 percent of the merchant's average monthly revenue. A business depositing $80,000 a month but maintaining an ADB of only $3,000 is a much higher risk than one maintaining $12,000.
• Deposit velocity and consistency: Are deposits coming in daily, weekly, or in large irregular chunks? A restaurant with daily small deposits is fundamentally different from a contractor getting three large payments a month. AI models are trained to understand these patterns by industry and flag abnormal deviations.
• Negative day frequency: How many days in each statement period did the account balance hit zero or below? One or two negative days in a month is manageable. Ten or more is a near-automatic decline at most shops. The model knows this because it has seen thousands of files with ten-plus negative days and knows how they performed.
• NSF and returned item count: Non-sufficient funds items are among the highest-weighted negative signals in AI underwriting. More than three to five NSFs per month, on average, will trigger an auto-decline at most modern funders. The model has learned that merchants with frequent NSFs struggle to sustain daily or weekly payment schedules.
• Existing MCA payment obligations: AI systems are increasingly good at identifying the signature of ACH debit patterns that look like MCA payments — small, regular debits that start right after a large deposit. They can estimate how many active positions the merchant is carrying and whether adding another one would breach a reasonable debt-to-revenue threshold.

Industry Benchmarks and Sector Modeling

One of the biggest advances in AI underwriting is the incorporation of industry-level performance data. A machine learning model trained on millions of MCA files doesn't just evaluate a merchant in isolation — it compares that merchant's cash flow profile to every other merchant in the same sector it has ever seen.

A restaurant with an average daily balance of $8,000 is evaluated differently than a staffing company with an average daily balance of $8,000. Revenue seasonality, typical payment behavior, and default rates vary dramatically by industry, and modern models price that risk into the offer automatically.

This is why two seemingly similar deals — same monthly revenue, same credit score, same time in business — can receive very different factor rates from the same funder. The industry variable is doing significant work in the pricing model.

Macroeconomic and External Signals

The most sophisticated underwriting engines layer in external data on top of the merchant's own financials. This can include sector-level economic indicators, local market conditions, and — increasingly — signals from the merchant's online presence. Review velocity, rating trends, and even social media activity have been incorporated into some funder models as early-warning indicators of business health.

These external signals carry less weight than direct cash flow data, but they matter at the margin, particularly for deals where the financial data alone is ambiguous.

Why AI Is Often Better for Merchants With Bad Credit

One genuinely positive development from AI underwriting is that it has reduced — though not eliminated — the weight placed on personal credit scores. Traditional bank underwriting is heavily FICO-dependent. MCA underwriting was always more cash-flow-focused, but human underwriters still used credit scores as a quick triage filter.

AI models trained on actual repayment outcomes have confirmed what experienced MCA underwriters have always suspected: cash flow is a better predictor of MCA repayment than personal credit score. A merchant with a 580 FICO but strong, consistent deposits and a healthy average daily balance is a better MCA risk than a merchant with a 720 FICO and an account that goes negative twice a month.

For brokers, this means the credit score conversation with merchants should shift. Stop leading with "your credit score is too low." Start asking about cash flow consistency, average daily balance, and NSF history. Those are the real gatekeepers now.

What This Means for Brokers: Practical Implications

Pre-Qualify on Cash Flow, Not Just Revenue

Revenue numbers are easy to get from a merchant. Cash flow quality takes a few more questions. Before you submit a file, ask your merchant:

• What is your average daily bank balance over the last three months?
• How many days in the last month did your account go negative?
• Do you have any current MCA positions? How many and what are the approximate daily payments?
• Have you had any returned items or NSFs in the last 90 days?

If the answers to those questions raise red flags, address them before submission — or set accurate expectations with the merchant about what kind of offer (if any) they're likely to receive.

Understand That Speed Is Not Random

When you submit to a funder with AI underwriting and get a decision back in under two minutes, that speed is a signal. An instant approval means the model found a clean profile — strong cash flow, no red flags, fits a known approval pattern. An instant decline means the model found a disqualifying signal quickly.

A slow response — several hours or a request for additional documents — usually means the file landed in a gray zone where the model's confidence was low and a human underwriter needs to make the final call. These are often deals that have one strong signal and one weak one: good revenue but poor average daily balance, for instance, or clean statements but a flagged industry.

Knowing this helps you manage the merchant relationship. Don't promise a decision by end of day on a file you know is borderline.

Statement Quality Matters More Than Ever

AI systems parse bank statements programmatically. Blurry PDFs, statements with broken formatting, or screenshots that are missing pages create parsing errors that either slow the process or cause a system to request resubmission. Submit clean, complete, properly formatted bank statement PDFs — ideally downloaded directly from the bank's portal, not photographed or printed and rescanned.

Some funders now offer bank login integrations (Plaid and similar platforms) that let them pull transaction data directly. When a funder offers this option, take it. It speeds up the process and eliminates the most common cause of underwriting delays.

Factor Rate Optimization Starts Before Submission

AI models price deals dynamically. The factor rate you receive is not a flat number that a funder assigned to your client's revenue tier — it's an output of the risk model, and it adjusts based on all the signals described above. You can influence that output at the margin by packaging files well.

For borderline files where the average daily balance is on the low side, consider asking the merchant to fund during a period when their balance is higher — typically after a strong sales week. For merchants with one or two older NSFs that don't reflect current behavior, a brief LOE (letter of explanation) submitted alongside the bank statements can help a human reviewer make the right call if the file escalates out of the AI queue.

How AI Is Changing Factor Rate Pricing

Dynamic pricing is one of the most significant ways AI has changed the MCA market. Before machine learning, factor rates were largely table-driven: here's our rate for 1-position merchants in this revenue band, here's our rate for 2-position merchants, and so on. Human underwriters applied those tables with some discretion.

Now, the best funders are pricing each deal based on a risk score that incorporates dozens of variables simultaneously. Two merchants with identical revenue and identical credit scores can receive different factor rates because their cash flow profiles are different, or because one is in a higher-default industry, or because one has a longer history of consistent deposits.

For brokers, this means you should stop assuming you know what rate a funder will offer before they run the numbers. Submit the file. You may be surprised — in either direction.

The Next Frontier: Predictive Renewal Offers

The most forward-looking funders are using AI not just to underwrite new deals, but to predict when existing merchants are ready for a renewal — and what offer they're likely to accept. By monitoring cash flow data in real time (with merchant permission, through bank integrations), some funders can generate proactive renewal offers before the merchant even pays down to the typical renewal threshold.

For brokers who have placed deals with these funders, this creates both an opportunity and a competitive risk. The opportunity: your merchant gets a seamless renewal experience and is more likely to stay with the funder (and by extension, your shop). The risk: if the funder has a direct sales channel, they may attempt to renew the merchant without going through you.

Understand the renewal policies of every funder in your panel. Some guarantee broker protection on renewals. Others don't. This matters more as AI makes renewal offers faster and more frequent.

Working With AI Underwriting: A Practical Takeaway

The shift to AI underwriting is not a threat to MCA brokers — it's a filter. Funders that have invested in machine learning can process more volume, price risk more accurately, and make faster decisions. That means more capacity for deals that fit their model and faster declines for deals that don't.

Your job as a broker hasn't fundamentally changed: match the right merchant to the right funder. What has changed is what "right fit" means. It used to mean matching revenue and credit score to a funder's published minimums. Now it means understanding which funders' AI models will score your merchant's cash flow profile favorably.

The brokers who thrive in this environment are the ones who do the cash flow analysis before submission, not after the decline. They know which funders use AI, which use human underwriters, and which use both. They understand that a fast decline is information, not a dead end — it tells you what that model found, which helps you find the funder whose model will see the deal differently.

Build your funder panel with underwriting methodology in mind. Know your funders not just by their published minimums, but by what their systems actually approve. That knowledge — accumulated file by file, approval by approval — is the competitive edge that no algorithm can replicate.