For QC firms, title agencies and lender QC teams

One hundred percent coveragefor ten files more than your mandated sample.

Taiso checks every checklist item on every loan file against the source document behind it — escalates only what it can't confirm to your QC reviewers, and produces exam-formatted evidence for the whole population. Same reviewers. Every file covered.

Scroll
The mandate is the floor, not the ceiling

The guideline says sample ten percent.
The repurchase letter doesn't care.

You review ten percent because that is what the guideline requires — and because reading a hundred percent by hand was never possible.
The defect that triggers a repurchase demand doesn't know it was supposed to land in your sample.
Every file outside the sample is a file you certified without opening. That exposure sits on your balance sheet, not the guideline's.
Adding reviewers raises your cost per file without raising your coverage above the mandated minimum.

Sampling was a concession to what people can read, not a judgment about what needs checking. Coverage is an architecture decision now.

Every loan, not the sample

One hundred percent of the population.
For ten files more than your sample.

Taiso reads every file, checks each checklist item against the source document behind it, and routes only what it can't confirm to your reviewers. Their hours stop going into re-reading and start going into findings.

Todayyou sample ten percent
in the mandated samplecertified unopened — your exposure

Everything you didn't open, you still certified.

With Taisoevery file, checked
checked against its source, certifiedthe few nobody could confirm → your reviewers
in the exam file — the record your examiner accepts

Ten files more than your sample. Every file systematically checked. The machine reads all of it; your reviewers decide the few it couldn't confirm — and the exam file covers the population, not the sample.

Start with Taiso Verify on a single workflow — no migration, and nothing new for your team to learn. Scale to the Taiso Agent Control Platform — your whole QC operation, verified.

What the mandate actually covers

The guideline sets a floor.
Your exposure sits above it.

10%
the mandated sample you review today
100%
checked against the source documents
90
files that used to be certified unopened

The mandate is satisfied either way. What changes is the ninety files you were already answering for.

See it run — one closed QC cycle

Watch every loan file get checked.
Not a sample.

A two-and-a-half-minute walkthrough of one hundred post-closing loan files — narrated by the QC lead who signs the exam report.

STEP 1 OF 6Your QC cycle today
NarratorOne hundred loan files land in post-closing QC today.
Post-closing · 100 loan files10 sampled · 90 unopened
Taiso verification gate — every checklist item checked against the source document in the file, before the cycle closes
in the mandated samplecertified unopenedchecked against sourcethe unconfirmed few
90files certified that nobody opened
10files your mandated sample actually covered

The ninety you didn't open are the ninety you still certified.

0of 100 checked against their source documents
confirmed0
the machine couldn't confirm0

No reviewer re-reads a file.

Loan file 29 of 100Income calculation — qualifying monthly income
The file qualified at$14,200
The source documents support“Year-to-date earnings and the verification of employment support $8,650.”
VerdictDISPUTED · held for a human
Certified clean, outside the sample. This is the defect class that becomes a repurchase demand two years later.
QC reviewer queue0
Your reviewers open these ten — not all one hundred. The same hours you spend on the mandated sample today, spent on the files that actually have findings.
Exam-formatted evidence0 / 100
confirmed0
decided by a QC reviewer0
certified unopened0

The record your examiner and your investor accept.

Sampled QC10 defects certified clean — surfaced at repurchase, if ever. Your reviewers re-read the sample by hand.
Full-coverage QC0 files certified unchecked · 10 decided by a reviewer · 100 in the exam file. Ten files more than your sample.
Re-run a closed QC cycle at full coverage →
←/→ step through · space plays · V mutes

Illustration of the mechanism on a hundred post-closing loan files. Volumes and counts are illustrative, not measured results — the exception rate most of all. How many items actually need a person is the number the parallel run measures on your own work, and it is the number that decides whether the hours hold.

What the gate checks

Five checks
on every checklist item.

01
Intent
Does the answer address the checklist item that was asked?
occupancy question answered from the appraisal → mismatch
02
Evidence
Is the fact in the file, or did the model infer it?
qualifying income located on the verification of employment → supported
03
Policy
Does the file follow the applicable guideline?
ratio above program maximum certified without exception → blocked
04
Risk
What does this cost you if it's wrong?
income calculation on a sold loan → a reviewer decides, not the model
05
Audit
Is the finding recorded, sourced and replayable?
item, source document, verdict, reviewer, timestamp → exam file

Five checks, one verdict — ALLOW — before the file is certified.

Before the agency signs

What the agency asks
before it accepts this.

BORROWER DATA
Loan files stay inside your boundary
Deploy in your own environment and nothing leaves it. On the hosted path it runs under a DPA with a no-training clause and retention you set.
THE MANDATE
Full coverage satisfies the sample, it doesn't replace it
Your mandated sample is still reviewed and still reported the way the guideline asks. Everything outside it is now checked too, in the same exam format.
YOUR REVIEWERS
Fewer files, every one with a finding
Reviewers stop reading ten files chosen at random and open the ones the machine could not confirm.

On certification: SOC 2 Type I is on the clock, and Type II is what most financial buyers gate on — we are not going to pretend otherwise. Closed work is still confidential data, so the pilot runs under NDA and a DPA, in your environment or ours, and goes through the same security review your clients require of any new processor. We will fill in the questionnaire.

What the exam file keeps

The exam file compounds.
The next audit starts ahead.

EXAM-FORMATTED EVIDENCE
A record your examiner accepts
Every checklist item, the source document behind it, the verdict, and the reviewer who decided it — formatted the way the exam asks for it.
YOUR CHECKLIST, EXECUTABLE
The guideline becomes the procedure
Your QC checklist stops living in a workbook and starts running as the thing that checks every file, the same way, every cycle.
COVERAGE YOU CAN STATE
A population, not a sample
When an investor asks what you reviewed, the answer stops being a percentage and starts being the file count.

By day 3 you see what it would have caught on a cycle you already closed. By day 14 you have the exam file and the list of what your sample missed — yours to keep either way.

Does this land?

What the exam
usually finds first.

A REPURCHASE LETTER ARRIVED
On a file outside the sample
The defect was real, the file was certified, and nobody had opened it. The guideline was satisfied and you still paid.
COVERAGE IS CAPPED BY HEADCOUNT
Ten percent is what you can read
Not what you would choose. Raising it means hiring reviewers against a fee that does not move.
THE INVESTOR ASKED
“What did you actually review?”
You have a sampling methodology. They wanted the population, the findings, and the evidence.

If two of these are true, the pilot below was built for you: a QC cycle you have already closed, and nothing in production changes.

The offer

One closed QC cycle.
At full coverage.

Give us a cycle you have already certified, under NDA. We re-run it at one hundred percent in parallel — no live file is touched, no investor is involved — and we hand back what the pipeline caught, what your sample missed, and what we got wrong. Bounded defect rates with the sample size beside them, never a zero-miss promise.