Reporting & Analytics

The numbers your board asks for, already assembled.

Funnel conversion, stage-level cycle time, branch and loan officer profitability, and continuous fair-lending monitoring — computed off the same ledger the operators work in, so nothing has to be reconciled.

Demo data · Quarter

Funded volume

$476,700,000

+22.1% YoY

Pull-through

91.6%

industry avg 74%

Cost per funded loan

$318

-83% vs. legacy stack

Avg. cycle time

9.2 days

34.1 days on prior LOS

Your ledger, right now

Computed live from the loans, leads and locks in this workspace

Loans in ledger

6 at closing or later

Pipeline volume

$14,899,947

sum of active loan amounts

Locked volume

0 commitments on the desk

Leads in CRM

33% of loans at closing or later

Lead-to-funded funnel

Conversion at every handoff

Leads captured

4,820—

Intake completed

2,91460%

Pre-approved

2,10772%

Application

1,63878%

Approved

1,42187%

Clear to close

1,33294%

Funded

1,30198%

Biggest leak is lead → intake at 60.5%. Agents now re-engage abandoned intakes within 90 seconds, which recovered 411 applications this period.

Cycle time by stage

Days, NexusMortgageOS vs. legacy baseline

Intake → Application0.4d

Legacy 4.2d

Application → Approval2.1d

Legacy 11.5d

Approval → CTC4.3d

Legacy 12.8d

CTC → Funded2.4d

Legacy 5.6d

Branch & channel P&L

Revenue net of fully loaded fulfillment cost

Channel	Units	Volume	Revenue	Cost	Net bps
Scottsdale · Retail	318	$121,400,000	$3,642,000	$1,014,000	216
Denver · Retail	264	$96,800,000	$2,904,000	$918,000	205
Wholesale · TPO	511	$187,300,000	$4,120,000	$1,186,000	157
Direct · Consumer	208	$71,200,000	$2,136,000	$902,000	173

Loan officer scorecard

Ranked by contribution

Dana Reyes
41 units
Pull-through 94%Cycle 8.4d238 bps
Marcus Hale
37 units
Pull-through 91%Cycle 9.1d221 bps
Priya Nandan
33 units
Pull-through 96%Cycle 7.8d244 bps
Owen Whitfield
28 units
Pull-through 83%Cycle 12.6d168 bps

HMDA & fair-lending monitoring

Continuous, not an annual scramble

LAR 100% field-complete

Approval rate parity98.4%

vs. control cohort, 90-day rolling

Pricing exception parity99.1%

Exceptions reviewed by second signer

Denial reason coverage100%

Every adverse action cites a rule, not a model

AI decision explainability100%

Model card + feature trace per decision

Exam-ready by construction

Every pricing exception, denial and AI recommendation carries the rule it fired on, the data it read and the human who signed it. LAR fields are validated at write time, so the filing is a export rather than a reconstruction.

· Quarterly disparate-impact regression across protected classes
· Adverse action notices generated within 24 hours, ECOA-mapped
· Model cards versioned with each agent deployment
· Immutable hash chain covering all 47 LAR data points