01 The beginning
At the desk
Loan origination system
- Amount financed
- $19,369.26
- Application
- App 7409
- Dealer
- Crestline Auto Outlet
- Borrower
- Travis Hutchins
Your last loan
App 7409 · Charged off$16,101.74 loss
Your next deal
App 9871 · ConditionalScore 34 → 62
For subprime auto lenders
Let your last loan inform your next deal.
Underbot connects your loan reports and underwriting, so what happened to your book informs the next deal.
In daily use. At a multi-state auto lender.
Two systems. One history.
One loan. The whole story.
Your loan origination system (LOS) holds what you knew when you bought the deal. Your loan management system (LMS) holds the payments and what followed.
Underbot
One connected history
02 The outcome
Down the road
Loan management system
- Loss
- $16,101.74
- Account
- 4036468
- Interest earned
- $976.40
Your reports show what happened; your underwriting puts those lessons to work on the deal in front of you.
See it in practice
From your book to the next decision.
Pick the question on your mind. Each short film shows the product at work.
01 · The dealer
Which dealers earn their volume?
Put the business a dealer sends beside what those loans became. See production, past-due balances and losses together before the next dealer conversation.
Explore Dealer Review02 · The loans behind the number
Which accounts need a closer look?
Start with the past-due total, then open the loans that make it up. Your team can see the accounts, balances and days behind that need attention.
Explore Collections03 · What came back
What did we actually recover?
Follow the dollars back to the accounts. See what has been collected and where the remaining recovery work stands.
Explore Recovery04 · The next deal
What should we do with this deal?
Apply your credit policy consistently, with the reasons in view. Bring the history of your book to the desk, while your underwriter keeps the final call.
Explore the Underwriting Engine05 · Proving the policy
Would our policy have caught the losses?
Replay past deals using only what was known when you bought them. Compare the calls with what happened, including the good loans you would want to keep.
Explore Policy BacktestThe films, stills and examples drawn on this page use fictional lenders, people and figures. Background photography, narration voice and music are AI-generated. Pressing play loads YouTube.
Part of the working day
At the desk. Across the book.
At a multi-state auto lender, the Underwriting Engine and the Dealer Review, Collections and Recovery reports are in daily use.
For the underwriting team
A consistent call. Clear reasons.
Deal 9871Revised structure · Tier D
Base score100
Credit history-20
Employment stability+28
Deal structure-14
Affordability+21
Collateral-8
Policy penalties-45
Engine call62ConditionalOut of 100
100 base + 7 factors + 0 bonuses - 45 penalties = 62
Maya Chen62 · Conditional
Alex Morgan62 · Conditional
Different underwriters. The same policy call.
Book history reaches the desk
App 7409 · $16,101.74 loss
Review context, separate from the score.The human makes the final call.Conditional · Stipulations required
For the reporting team
The number. The loans behind it.
Crestline Auto OutletDealer review
Past-due balance · 30+ days$915,659.90
78 accounts. One mark per loan.App 7848: 1.31%
75 other accounts$872,726.54
From the charged-off book
App 7409 · $16,101.74 loss
The loan above. Outside the past-due total.Follow a number to its loans.
Bring what happened into the next decision.
One connected history. A better-informed next decision.Illustrative product excerpts. Fictional loans and underwriters.
Fits the way you work
Keep your systems. Keep the final call.
Start small. One sheet, ten deals, no names. Back in three business days.
Inside your building
Nothing identifying leaves
The review sheet has no box for a name, a Social Security number, a VIN or a dealer. Files, exports and connections wait for a signed NDA.
One-way access
Read-only access
Underbot reads your records. It never writes to your systems of record.
Across the boundary
Encrypted in transit and at rest
Your files are encrypted from the moment they arrive.
Your work stays yours
No AI training on your data
Your data is never used to train or fine-tune AI models.
A memo comes back. The final call stays with you.
Nothing needs replacing. Underbot can use a connection your system already provides, called an API, read-only access to your records, or the export you already produce.
Read about data handling
Your next step
Free 10-deal review
See what your last ten deals can teach you about the next one.
How it works
Get the sheet.
Fill in ten rows.
Any ten finished deals. The best mix: a few that paid, a few that went bad, a few you declined and one your team argued about.
No names. No Social Security numbers. No dealer. The sheet asks about the deal, never about who.
Email it to hello@underbot.ai.
Your memo comes back within three business days.
- 01Ten deals re-scored on one standard
- 02The math behind every score
- 03A plain-English findings memo
- 04A 30-minute walkthrough, if you want one
If we can't show you at least one thing about your own book you didn't already know, we'll tell you straight.
The sheet has no box for a name, an address, a Social Security number, a birth date, a VIN or a dealer. Want an NDA first? Ask and it's in your inbox the same day.
See a sample findings memoA look at the findings memo
01 / Ten deals re-scored
Ten deals.
One clear standard.
Each deal gets a score, a decision and the reasons behind it. Here is one example from the films.
10 files. The same policy. Every reason in view.
A closer look App 9871
- Original structure
- 34Decline
- With $1,500 more down
- 62Conditional
A different structure. A different decision.
02 / The math behind every score
Every point on the page.
Deal 3 scores 100 + 67 + 15 - 45 = 137. Tier A, Approve.
Helps most: job tenure of 5+ years (+20), DTI under 35% (+15), owns home (+12).
Hurts most: LTV of 110 to 119% (-10), a 72-month term (-10), one or two 30-day lates in 12 months (-8).
It passes, but the risk stack is hot enough to review before funding.
03 / A plain-English findings memo
A change worth considering.
The original structure falls short of the policy. More cash down brings it into the conditional tier.
Confirm the borrower can meet the terms and complete the stipulations before making the call.
Your team keeps the final call.
04 / The 30-minute walkthrough
Make sense of it.
Together.
- Review the ten decisions and their reasons.
- Go through the flags and your open questions.
- Agree what, if anything, your team would change.