OnDeck Just Killed the Paperwork: What 'A Decision in Under a Minute' Means for Borrowers

What OnDeck actually announced

On October 8, 2026, OnDeck (owned by Enova) announced it had automated its final decisioning: machine-learning underwriting models combined with AI-enabled verification, replacing the traditional stipulation process for qualifying files. The headline number from the announcement: a stipulation-free decision in under a minute. The company says its goal is to make most approved applications stipulation-free offers, and the system is rolling out to new customers across all channels after a pilot with existing customers.

Jim Granat, OnDeck's Co-President of Small Business Lending, framed it as a customer-experience play rather than a speed play: "We didn't set out to be the fastest lender – we set out to be the easiest." The operational promise: qualified applicants "can receive approval with no stipulations, allowing them to move directly to funding without the need to submit additional documents."

Read that promise carefully, because the important words are the boring ones. Not "decision in under a minute" — that is the headline. The important words are "without the need to submit additional documents." That is the thing that actually kills most applications.

What "stipulation-free" means in plain English

A "stipulation" is lender-speak for a document condition attached to your approval. In the traditional flow, you apply, an underwriter reviews your file, and then you get a list: send three months of bank statements, send the voided check, confirm this deposit, explain this overdraft. Every round of stipulations adds a day or two — and every round is another chance for the deal to fall apart because a busy owner forgets to reply to the email. Our MCA approval-timeline guide walks through the old version of this dance; the stipulation chase is usually the slowest part of it.

Stipulation-free means the verification that used to come from a human underwriter reading your PDFs now comes from automated systems: identity verification, bank-data pulls, and models trained on enough applications to spot a clean file at machine speed. When the system can verify everything itself, there is nothing to stipulate. The application goes straight from decision to funding.

Two things this does not mean. First, "decision" is not "funding" — you still accept terms and the money still moves through the usual rails, which takes hours, not seconds. Second, this applies to qualifying applications, not all applications. The operative word in OnDeck's announcement is "qualified." If the automated systems cannot verify something cleanly, a human still gets involved and the stipulations come back.

Why this is a trend, not a one-off

OnDeck did not invent this direction; it is the latest big name to commit to it publicly. Across the industry, the borrower-facing application is converging on the same sequence: capture your information, verify your identity automatically, pull your bank data directly instead of asking you to upload PDFs, extract the underwriting signals from that data, and route only the ambiguous cases to humans. Fundbox's much-reviewed 3-minute application (a $250,000 line of credit with a 600 minimum credit score and just three months in business, per WSJ Buy Side's recent review) is the same philosophy: shrink the paperwork surface until there is almost nothing left for the borrower to do.

The economics are obvious once you say them out loud. Every document request is a drop-off point. Industry data has shown for years that funding applications die on friction far more often than on credit: the owner who never sends the second round of statements, the voided check that never gets uploaded, the processor statement buried in an email thread. The lenders that figured out how to verify without asking are winning the files — and everyone else is now racing to copy the playbook.

For borrowers, the practical consequence is simple: the application itself is becoming the smallest part of getting funded, and the quality of the data behind it is becoming the largest part. When verification is automated, there is nowhere to hide a thin file. Which brings us to the catch.

The catch: machines decline at machine speed too

Automated decisioning is symmetric. It approves clean files in under a minute — and it rejects messy ones in under a minute, too. The same system that eliminates the stipulation list also eliminates the human underwriter who might have looked at your file, noticed the seasonal dip in your deposits, and asked you about it instead of declining.

What the models see is what matters, and what they see comes almost entirely from your bank data: deposit volume, average daily balances, negative-balance days, existing advance positions and leverage, chargebacks, and industry risk. We break down exactly what underwriters read in what lenders see in bank statements — and in an automated world, that checklist is not just what the underwriter reads, it is quite literally the application.

This is why "apply everywhere and see who says yes" is bad advice now. Every automated decline is a data point, and merchants with stacked positions, undisclosed advances, or cash-flow gaps that do not show up well in a bank-data pull are the files automation is least forgiving toward. Comparing the true cost of offers still matters — but only after you survive the machine that decides whether you get an offer at all.

What to do before you apply

The new rule of fast funding is: prepare for the machine, not the human. A few practical moves before you hit submit:

• Apply with complete, consistent information. The #1 trigger for a stipulation list — or a fast decline — is data that does not reconcile: a revenue figure on the application that does not match the deposits, an address that does not match the business records. Automation checks all of it in seconds.
• Do not hide existing positions. If you have an outstanding advance, the automated verification will see the daily debits in your bank data. Undisclosed positions discovered by the machine read as a red flag; disclosed positions are just math. If you are stacked, understand your consolidation options before you apply for more.
• Time your application. Automated systems read recent bank activity. Applying right after a strong deposit month — rather than right after payroll drained the account — changes what the model sees. Details matter when the reader is an algorithm.
• Have your statements ready anyway. "Stipulation-free for qualifying applications" still leaves the unqualified ones. Keep three months of statements and your voided check handy — the old flow in what happens after you upload bank statements is still the fallback for files the machine cannot verify on its own.
• Read the actual offer, not just the approval. A one-minute decision can still be an expensive one. Compare total payback across offers before you sign — speed of approval says nothing about cost of capital.

The broker's edge in an automated world

Here is the part no lender's press release will tell you: automated decisioning made the single-lender application more of a lottery, not less. One lender's model is one opinion — one yes-or-no verdict from one algorithm trained on one portfolio. If its model does not like your industry, your deposit pattern, or your seasonality, you get a machine-speed no and learn nothing.

A broker's job in this world is file preparation at scale. We pre-build your file — complete data, reconciled numbers, the right product for the right funder — and run it across 75+ lenders and funders, each with its own automated decisioning. The same file that one lender's machine declines in under a minute is another lender's machine's ideal approval. That is not a theory; it is what we see every day across dozens of different underwriting models. Our direct-lender-vs-broker breakdown covers the structural reasons in more detail.

The industry's direction is clear: the paperwork is disappearing, and the quality of the file behind the application is everything. OnDeck is just saying it out loud. Start an application — it takes about five minutes, does not affect your credit score, and we will shop it across the full market instead of one algorithm.

Frequently asked questions

What did OnDeck announce on October 8, 2026?

OnDeck (Enova) announced automated final-decisioning that combines machine learning with AI-enabled verification, delivering a stipulation-free decision in under a minute for qualifying applications. It rolled out to new customers across channels after a pilot with existing customers, with the stated goal of making most approved applications stipulation-free.

What does "stipulation-free" mean?

It means no follow-up document requests: the approval does not come with a list of stipulations (extra statements, explanations, voided checks). Automated systems verify what a human underwriter used to check manually, so a qualifying application moves directly from decision to funding.

Does a one-minute decision mean one-minute funding?

No. The decision is fast; funding still takes the usual hours. You also still have to accept the terms. Speed of approval says nothing about the cost of the offer — compare total payback before signing.

Will every application get a stipulation-free decision?

No. The operative word is "qualifying." If the automated systems cannot verify something cleanly — inconsistent data, thin bank history, undisclosed positions — the file goes to a human and the traditional stipulation process returns.

How should I prepare for automated underwriting?

Make your data reconcile: the revenue on your application should match your deposits, disclose existing advance positions (the machine sees the debits anyway), apply after a strong deposit month if you can, and keep three months of statements handy as a fallback. Then have someone who shops the full market run the file, because one algorithm's no is not the market's no.

Is AI replacing brokers in business funding?

AI is replacing the paperwork, not the broker. Automation made the application smaller and the file behind it more important — and it made each lender's decision a single algorithmic verdict. Someone still has to decide which funders' algorithms your file fits best, and that is a broker's job. We wrote more about the limits of AI in funding in what AI brokers can and cannot do.