Documents in lending

AI Underwriter

Every file the same rigor. Every memo drafted in seconds. Every decision yours to sign.

Spreads the financials, reads the story behind the numbers, and drafts the memo a senior analyst would write — calibrated to your credit policy. Your team edits and signs.

What it does
Spread · Story · Memo
Source citations
Every claim
Tunable to
Your credit policy
Final decision
Your underwriter
Used across
SME & commercial lending·CDFI underwriting

A defensible memo, drafted

A defensible memo, before your team sits down.

Kita's AI Underwriter reads every page of every document, models cash flow, and drafts a credit assessment with cited evidence — calibrated to your credit policy. Every number traces back to a source line. Nothing is invented.

Applicant

Verde Logística S.A. de C.V.
Mexico City · Trucking · 36 employees
Requested MXN 16M
Term 36 mo
Purpose Fleet expansion (12 trucks)
Recommendation · APPROVE
MXN 14.4M · 36 mo · 16.5% APR

Approve below requested. Cash flow supports debt service at 1.42× DSCR over the trailing 12 months. Counter at 80% of ask given concentration in two contract counterparties (61% of receipts).

Cash flow

Trailing 12-mo operating cash flow MXN 8.1M / yr. Margin holding at 14.6% across 73 monthly statements reviewed across BBVA and Santander.

Debt service

Existing obligations MXN 1.6M / yr. Pro-forma debt service with this facility MXN 5.7M / yr, leaving DSCR 1.42× at base case, 1.19× at -15% revenue stress.

Risk notes

Counterparty concentration: top two clients drive 61% of receipts. One late SAT filing in Q2 2025, since cured. Tax compliance current per constancia de situación fiscal.

Drafted by Kita AI Underwriter · 17 sources · 2.4s

Policy check
SME Term Loan · MX-2026-Q2
5 / 6 pass
DSCR ≥ 1.25 1.42
30-day arrears = 0 Yes
Tax compliance current Yes
Operating tenure ≥ 24 mo 8.2 yr
Margin ≥ 8% 14.6%
Concentration < 50% 61%
Final approval is always a human at your institution.

The problem

Credit assessment is uneven.

  1. Inconsistent decisions
    Different analysts apply different standards. Portfolio risk compounds when underwriting quality varies across your team.

  2. Generic models don't fit your book
    Off-the-shelf credit scores weren't built for your borrower profile. They miss the signals that actually predict repayment in your portfolio.

  3. No feedback loop
    You generate outcomes data with every loan, but your underwriting logic never learns from it. The same blind spots repeat.

  4. Speed vs. quality tradeoff
    Scaling volume means either hiring more analysts or accepting lower-quality decisions. Neither is sustainable.

What it does

An analyst, in your stack.

Not a black-box score. Kita does the same three things a senior credit analyst does — spreads the financials, reads the story behind the numbers, and drafts the memo. Faster, and at scale.

i. Financial spreading
Pulls statements, GLs, and bank data into a normalized period-over-period spread. DSCR, margin, debt service, trend — all computed in one pass.

ii. Story behind the numbers
Reads the qualitative narrative the spread alone won't tell you — why margins moved, what concentration risk looks like, what an arrears spike actually signals.

iii. Drafts the memo
Produces the credit memo a senior analyst would write — recommendation, evidence, risk notes — every claim cited to a source line. Your team edits, signs, sends.

Capabilities

More than a model.

Kita does the work an analyst does — spreads the financials, reads the story behind them, and drafts the memo. Calibrated to your credit policy. Every number traces back to a source line. Nothing is invented.

  • Financial spreading — Pulls financials and bank data into a clean period-over-period spread. Ratios, trend, debt service modeled out of the box.
  • Reads the narrative — Picks up the qualitative story behind the numbers — concentration risk, seasonality, why margins moved, what late filings imply.
  • Drafts the memo — Produces the credit memo your senior underwriter would write. Recommendation, evidence, risk notes, ready to edit.
  • Cited to source — Every claim in the assessment links back to the document line that produced it. Audit in one click.
  • Calibrated to your policy — Tune the decisioning criteria that matter to your book. Not locked into a generic credit score.
  • Adaptive — Recommendation quality improves loan cycle over loan cycle as your portfolio grows.

Why it matters

  1. Designed for volume
    Hundreds or hundreds of thousands of applications per month. Scales without adding operational overhead.

  2. Outcome-driven
    Traditional scorecards are static. Kita learns from what actually happens in your portfolio.

  3. Wires into your stack
    API-first. Integrates into your existing LOS, credit engine, or custom workflow.

Underwrite more borrowers. Faster. With cleaner data.