How OnDeck's SMB Growth Report Reveals What AI Underwriting for Merchant Cash Advance Must Deliver

Small Businesses Are Adopting AI Faster Than Their Lenders

OnDeck's latest small business report, released in August 2026, carries a headline that should concern every MCA funder still relying on manual underwriting: small businesses are leaning into growth and AI at a pace that outstrips the lenders serving them. The report, drawn from OnDeck's survey of thousands of SMB owners, found that confidence in growth has climbed even as economic uncertainty lingers, and that AI tool adoption among small businesses has accelerated meaningfully over the past year.

This matters for AI underwriting for merchant cash advance because the merchants applying for funding are no longer the same merchants funders built their models around. A restaurant owner using AI to forecast inventory needs generates different deposit patterns than one managing by gut. A contractor using AI scheduling tools smooths out the revenue volatility that traditional underwriting flags as risk. The cash flow signatures are changing, and underwriting systems that rely on static rules or manual review are increasingly misreading the signals.

At the same time, the 2026 Inc 5000 list spotlights funders growing at staggering rates: Specialty Capital at nearly 2,000% three-year growth, Parafin at 969%, FundCanna carving out a niche vertical. The funders on that list did not achieve those numbers by reviewing bank statements in spreadsheets. They built, bought, or integrated systems that let them underwrite faster and more accurately than their competitors. The question for everyone else is straightforward: what does your underwriting stack need to deliver to stay in the conversation?

What AI Underwriting for MCA Must Actually Do Now

Reading Cash Flow That Merchants Have Already Optimized

The traditional MCA underwriting model is built on a simple premise: bank statements tell the truth about a business. Monthly deposits indicate revenue. Daily balances reveal stability. NSF counts signal distress. These remain useful signals, but they are no longer sufficient when the merchant on the other side of the application is using AI tools to manage their own finances.

Consider what happens when a merchant adopts AI-driven invoicing that accelerates collections. Their average daily balance rises, but not because the business fundamentally changed. Or consider a merchant using dynamic pricing software that smooths seasonal revenue dips. The bank statements look healthier, but the underlying business risk profile is identical to last year's. AI underwriting models need to distinguish between genuine improvement and optimization artifacts. That requires pattern recognition across transaction categories, not just top-line deposit totals.

Purpose-built document analysis models handle this by categorizing individual transactions, flagging unusual deposit clustering, and comparing intra-month patterns against historical baselines. General-purpose LLMs struggle here because they lack the domain-specific training data to know what a "normal" deposit pattern looks like for a landscaping company versus a dental practice. As we explored in our analysis of how cash flow data depth separates winning MCA underwriting from guesswork, the funders pulling ahead are the ones ingesting more granular data, not more documents.

Speed Is Now an Underwriting Signal, Not Just a Sales Metric

MCA has always been a speed game. The broker who gets the offer out first wins the deal. But OnDeck's report highlights something subtler: SMBs increasingly expect the funding process itself to feel like the AI-powered tools they use everywhere else. Instant. Frictionless. Mobile.

This expectation creates a feedback loop. Merchants who experience slow, clunky application processes self-select out, leaving funders with a skewed applicant pool. The best merchants, the ones with strong revenue, clean books, and multiple funding options, go to the platform that responds in minutes, not days. Speed becomes a risk filter. Slow processes attract desperate applicants.

Async document collection directly addresses this. When a merchant can upload bank statements from their phone in under two minutes, the funder gets the documents faster and the merchant stays engaged. Let's Submit's upload links, for example, let merchants drag and drop PDFs, snap photos of statements, and submit signed applications from a single mobile-friendly page. The AI extraction layer then parses revenue, deposits, daily balances, and NSFs automatically, so underwriters open a clean, structured application rather than a pile of raw files.

The result is that the funder's speed-to-offer compresses from days to hours. And as the Inc 5000 fastest-growing MCA funders demonstrate, that speed advantage compounds over hundreds or thousands of deals per month.

Fraud Detection That Scales With Volume

Growth creates a paradox for MCA funders: the more deals you process, the more surface area you expose to fraud, and the less time any individual underwriter has to scrutinize each file. Manual fraud detection does not scale linearly. It degrades.

AI fraud detection for business lending addresses this by running every document through pattern-matching models trained on known fabrication techniques. Font inconsistencies, pixel-level metadata anomalies, transaction sequences that violate banking system norms, deposit amounts that land on suspiciously round numbers. These checks happen in seconds, before a human underwriter ever sees the file.

The Parafin model is instructive here. At 969% three-year growth, they are processing an enormous volume of advances through their platform integrations. That volume is only sustainable if fraud detection is automated and embedded in the intake pipeline, not bolted on as an afterthought. Independent funders competing for the same merchants need the same capability, even without the platform data advantage that Parafin enjoys.

The Platform Lender Squeeze and What It Means for Independents

OnDeck's growth confidence findings do not exist in a vacuum. Shopify originated $1.4 billion in small business loans and MCAs in Q2 alone. Square's loan cohorts continue posting sub-4% loss rates. These platform lenders underwrite on transaction data that independent funders simply cannot access. They see every sale, every refund, every chargeback in real time.

Independent MCA funders will never have that data advantage. But they can close the gap on two fronts: document intake speed and analytical depth on the data they do collect. Bank statements remain the primary underwriting artifact for independent funders, and the quality of analysis applied to those statements determines whether a funder's loss rates stay manageable at scale.

This is where the build-versus-buy decision becomes critical. In 2026, the cost of assembling an AI-powered document analysis pipeline from open-source components has fallen dramatically. But maintenance, model retraining, edge case handling, and compliance logging still consume engineering bandwidth that most funders would rather spend on origination. Platforms like Let's Submit offer a middle path: managed AI extraction and async collection without the overhead of building a proprietary system.

The competitive landscape also explains why funders on the Inc 5000 list tend to cluster in two categories. Either they are platform-embedded lenders with proprietary data moats, or they are independent funders who invested early in automated verification infrastructure. The middle, funders with high volume but manual processes, is where margins collapse and losses spike.

Frequently Asked Questions

What is AI underwriting for merchant cash advance?

AI underwriting for merchant cash advance refers to using machine learning models and automated document analysis to evaluate a merchant's funding application. Instead of relying solely on manual review of bank statements, AI systems parse transaction data, categorize deposits, detect anomalies, and flag fraud indicators automatically. This allows funders to make faster, more consistent decisions while reducing the risk of human error. The technology is particularly valuable at scale, where manual underwriting creates throughput bottlenecks.

How does SMB AI adoption affect MCA underwriting?

When small businesses adopt AI tools for invoicing, pricing, or cash management, their bank statement patterns change in ways that traditional underwriting rules may misinterpret. Revenue smoothing, accelerated collections, and optimized payment timing can make a business appear healthier or riskier than it actually is. MCA underwriters need models sophisticated enough to distinguish between genuine business improvement and artifacts of financial optimization software.

Why does async document collection matter for MCA funders?

Async document collection lets merchants submit bank statements, IDs, and signed applications on their own time, from any device, without needing to coordinate a live call or in-person meeting. This matters because the best-qualified merchants, the ones with multiple funding options, choose the path of least resistance. A funder who sends a mobile upload link and receives documents in minutes has a structural advantage over one who emails back and forth for days. Async collection also feeds directly into AI extraction pipelines, eliminating the manual data entry step entirely.

Can independent MCA funders compete with platform lenders on underwriting speed?

Independent funders cannot match the proprietary transaction data that platform lenders like Shopify and Square access. However, they can compete on underwriting speed by automating the document intake and analysis steps that currently consume the most time. By combining async collection tools with AI-powered bank statement extraction, independent funders can compress their time-to-offer from days to hours. The latest origination numbers from major lenders confirm that this level of automation is now expected, not optional.

Conclusion

OnDeck's 2026 report confirms what the Inc 5000 list already proved: small businesses are moving faster, adopting AI tools, and expecting their lenders to keep up. MCA funders who still rely on manual bank statement review and email-based document collection are falling behind on speed, accuracy, and fraud detection simultaneously. AI underwriting for merchant cash advance is the baseline, not the differentiator.

The funders winning right now combine automated document analysis with frictionless merchant-facing intake. Let's Submit brings both together: AI-powered extraction that pulls revenue, deposits, and risk signals from bank statements automatically, paired with async upload links that merchants complete in minutes from their phones. Visit letssubmit.ca to see how async verification and AI extraction fit into your underwriting workflow.