A representative submission shape. Five standard controls passed it cleanly. Ultiwrite read the same file and found a business that does not add up.
Checklist verdict: clear to board.
Five controls said yes by checking the merchant against a list. Ultiwrite said wait by checking the merchant against itself. That is the entire difference — and it is the difference between a write-off and a decline.
Does this merchant's story hold together? Each layer interrogates a different part of it. A lone weak signal is noise; a cluster that reinforces across layers is a pattern no isolated check can see.
A real principal accretes a decade of incidental digital exhaust no one bothered to fabricate — because no one knew to. We weigh footprint depth, account age, and breach-history corroboration. A frictionless, brand-new identity is not reassuring. It is the tell.
Legal name, designator, formation date, state, EIN, and tax-filing name are checked against each other before they are checked against anything else. An entity formed last month behind an aged domain is not established — it is borrowing standing it did not earn.
Domain age is cheap to buy; a coherent operating footprint is not. We measure shared infrastructure, reachable contact points, and the distance between a confident storefront and a hollow back office.
We do not ask whether a field is filled in. We read paperwork for internal consistency and the structural fingerprints of files that were altered, assembled, or recycled out of someone else's submission.
Six behavioral vectors — timing, card mix, MCC behavior, revenue flow, structure, multi-processor stacking — fold into one Statistical Deviation Score with a confidence interval. One anomaly is noise. A cluster that corroborates is a pattern.
The widening gap between the business a merchant described at boarding and the business its transactions describe over time. The slope of that divergence bends upward ahead of the first chargeback — the window in which intervention is still cheap.
Review, underwriting, monitoring, and portfolio intelligence — four ways to see the merchant behind the application, not just the form.
Ultiwrite scores whether a merchant's identity, entity, owners, web presence, documents, and behavior all tell the same story — surfacing incoherence a form review misses.
An underwriter's question is not only whether to approve — it is how much to reserve, and what it costs to be wrong. So every read resolves into a figure. An acquirer's exposure is not the merchant's monthly volume; it is the outstanding chargeback tail at the moment the merchant stops answering the phone — and when payment precedes delivery, that clock starts at expected delivery rather than at the sale. Drive the inputs and watch the figure separate from the naive read.
Payment precedes delivery, so the dispute clock starts at expected delivery — not at the transaction.
$150k/mo × 180days ÷ 30
180 days = 60 fulfillment lag + 120 dispute window
A volume-based read understates this merchant by 6.0×. That gap is reserve that was never held.
EL = PD × LGD × EAD
Outstanding chargeback tail the moment the merchant stops answering.
No recovery assumed. Zero is the honest default until an acquirer supplies a realized rate.
Fitted to your resolved outcomes. Portfolios differ too much for another acquirer's default rate to be anything but noise on yours.
Exposure and severity resolve on day one. Both compute from fields already collected at intake — no integration, no historical data, no waiting. Probability then calibrates against your own resolved outcomes as they land, which is the only way a default rate means anything on your portfolio. Most tools invert this: a black-box probability, and no exposure figure at all.
Exposure arithmetic is computed from the inputs above. The 120-day dispute window and zero-recovery default are priors — stated, conservative, and tunable per program, because a number you cannot audit is a number you cannot underwrite on. Long-dated travel and events run materially past 120 days, so this is the conservative floor for ordinary merchants, not the ceiling.
Every merchant is analyzed across identity, entity, owners, web presence, documents, and processing history — cross-referenced for a single coherence score.
Identity, entity, owners, web presence, documents, and processing history — cross-referenced to detect incoherence.
Ultiwrite runs server-to-server and plugs into the CRMs, data warehouses, document stores, and alerting your underwriting and boarding teams already use.
Ultiwrite surfaces coherence signals, reason codes, and recommendations. It never makes the boarding decision — your underwriter always does.
Never renders an autonomous approve or decline.
Recommendations only — your underwriter decides.
Every signal is explained and inspectable.
Supplements — does not replace — compliance.
A server-to-server v1 API for scoring, review, monitoring, and portfolio analysis. Every endpoint runs in shadow mode and returns reason codes, never autonomous decisions.
Bearer uw_live_ keys issued to pilot partners.
Score one application or upload a portfolio CSV.
Structured, inspectable signals on every response.
Discoverable at /.well-known/openapi.json.
Ultiwrite catches the merchants whose story doesn't add up — the ones a clean-looking application would sail right past.
Risk Lead
Underwriting, ISO Partner
For small ISOs or early pilot users
For active ISOs and underwriters
For teams monitoring boarded merchants
Ultiwrite is provisioned manually for approved ISOs, PayFacs, processors, acquirers, sponsor banks, and underwriting teams. Start with a pilot.
Access is reviewed manually — submitting a request does not guarantee access.
