Every analyst underwrites differently, and I can’t tell how well my policy is applied.

What it looks like from the credit desk.

1. The same file gets two answers depending on who reviews it.

   The policy is written down, but each analyst reads it their own way, and nobody measures the gap.

2. Underwriting a small business takes days.

   Bureau report, SAT tax compliance opinion, tax records and bank statements show up through different channels, and someone stitches them together by hand.

3. The policy lives in a manual and in two people’s heads.

   When one of them is out, the committee stalls or decides without the full picture.

4. When the committee overrules, nobody writes down why.

   Months later, in front of a funder or an auditor, no one can reconstruct the exception.

The agent

A credit engine wired into your sources: the bureau, the SAT and the system where your portfolio lives. It takes the application, builds the file, applies the policy and drafts the memo. It flags when a channel stops sending applications and when it can’t do its own job.

The accountable person

Your head of credit owns the engine, by name. For the first 4 weeks, the engine checks with a person on every decision and action. After that, anything high-stakes —every approval, every amount— always goes through a person; simple things, like asking for a missing document, it handles with what it learned, and every one is logged.

The evidence

Every memo records which sources it pulled and when, which rule it applied and what the person decided. It’s auditable: in front of your funder, your auditor or your regulator, the file explains every decision and every exception.

The stop button

Your policy’s hard rules are written in code and covered by automated tests: if a test fails, the change doesn’t ship. Your team can pause the engine at any time and fall back to the manual process.

A policy you can’t prove doesn’t help you with a funder.

What a lender asks before starting.

Can an AI credit engine approve loans on its own?

Not the way we build it. The engine gathers the data, applies your policy’s criteria and hands back a memo that explains its recommendation. Your committee or leadership signs off on approval and amount, and can decide differently with a note that stays in the file. Every decision keeps an owner, and every exception is written down.

What data does the engine use to underwrite a small business?

Whatever your policy already requires: the credit bureau report, the SAT tax compliance opinion and the tax file, plus anything your process adds, such as bank statements. The engine doesn’t make up criteria. It applies the ones your credit team wrote, in the same order for every application, and records each source it pulled and when.

How do I know the engine applies my policy correctly?

Every hard rule is written in code and has an automated test that checks it before each change; if one fails, the change doesn’t ship. Every memo also states which rule it applied. In the first weeks, your team compares memos against your analysts’ calls and tightens the policy wherever they disagree.

What does it cost, and how do we start?

You start with a free diagnostic: we review your credit policy, your data sources and your application flow, and tell you which parts can live in an engine and which should stay with your analysts. Pricing comes in a formal proposal. There is a one-time activation fee and a monthly operations fee, because InnovaBlack keeps the engine supervised and current.