Chris Wyatt
Venture Concept · Commercial Card Underwriting Infrastructure
Essay

Pulse Underwriting

Card networks and issuers cannot underwrite long-tail B2B suppliers because they cannot see eighteen months of bank statements. Pulse is the real-time underwriting layer that turns live AR/AP signal into a synthetic credit file, unlocking commercial card acceptance for the suppliers most B2B platforms reject today.

$89T
Global B2B payment flow, 2024
Juniper Research
$575B
US commercial card purchase volume, 2023
Nilson Report
$1.6T
B2B working capital trapped in payable terms
PYMNTS-American Express

01 · The problem

Half the B2B supplier base is invisible to commercial-card underwriting

Visa, Mastercard, and the major issuers can profitably issue commercial cards to suppliers with eighteen months of clean bank data. The other half of the supplier universe (newer entities, sole proprietors, complex ownership structures) cannot get underwritten at all. The result is the largest under-served credit market in B2B.

18 months
Bank-statement requirement for traditional issuance
Industry standard
9-12 days
Median manual underwriting cycle for SMB issuance
Operator estimate
60%+
Of small US suppliers cite credit access as primary growth barrier
Federal Reserve SBCS, 2024

02 · The thesis

Live AR/AP signal beats eighteen months of bank statements

A supplier's AR aging, payment-terms history, and customer concentration tell you more about their next-90-days credit than eighteen months of bank statements ever did. The data exists, scattered across ERPs, AP platforms, and bank feeds.

Pulse aggregates that signal, normalizes it, and emits a real-time underwriting score issuers can plug into their existing flow. Synthesis is what's missing. Modeling capacity is not.

03 · The product

What it does

01

Live AR/AP signal aggregation

Pull from ERPs, AP platforms, bank feeds, and B2B network metadata into a unified credit profile.

02

Synthetic underwriting score

A real-time, explainable score callable from any issuer's underwriting flow, with feature-level provenance.

03

Continuous re-underwriting

Credit lines adjust on AR aging signal, not on annual review cycles. Defaults compress.

04

Embedded finance API

Drop-in endpoint for B2B platforms that want to extend payment terms to suppliers their card processor would otherwise reject.

04 · Why now

The timing case

  1. 1

    Visa and Mastercard have publicly committed to expanding commercial-card acceptance into the long tail. The underwriting bottleneck is the constraint.

  2. 2

    Open banking and ERP API access are now production-grade in the US, removing the data-acquisition cost that historically made this uneconomic.

  3. 3

    Embedded-finance infrastructure has matured. B2B platforms now expect to issue credit themselves rather than route customers to a third-party issuer.

  4. 4

    Visa Commercial Solutions and Mastercard Track both published long-tail commercial card commitments in 2024, targeting the 60 percent of US B2B spend still on check and ACH. Card-acceptance economics now favor issuers who can underwrite suppliers in days, not weeks. The Federal Reserve's 2024 Small Business Credit Survey found 60 percent of small suppliers cite credit access as a binding growth constraint. Codat and Rutter are commercially available for ERP and accounting API access in the same window. The data layer matured. The underwriting layer did not.

05 · Why I see it

The view from inside the work

I built Finexio's supplier enablement engine across a multi-bank stack and have spent a decade on the issuer-network side of the market. The constraint is not credit appetite. It is underwriting input.

06 · Comparable references

What's already in the market, and where the gap is

An honest read on the adjacent landscape. Not every comparable is a competitor. Some are partners. Some are the market the venture displaces.

Reference
Dun & Bradstreet, Experian Business
What they do
Static commercial credit bureaus built on tradeline reporting.
Gap
Backward-looking. Pulse uses live AR/AP signal that updates daily, not quarterly.
Reference
Plaid, Codat, Rutter
What they do
Financial-data aggregation across bank and accounting systems.
Gap
Sells the data feed. Pulse sells the underwriting score on top of it.
Reference
Resolve, Balance, Slope
What they do
Embedded B2B trade-credit products with proprietary underwriting.
Gap
Bundle the underwriting with the credit balance sheet. Pulse is balance-sheet-agnostic and underwrites for any issuer.
Reference
Issuer-internal underwriting
What they do
Manual review on bank statements and tradelines, 12-day cycles.
Gap
The status quo Pulse compresses. Two weeks becomes two minutes.

07 · Key risks

What could break the thesis

Operator-grade pre-mortem. Surfaced because the buyers and partners worth talking to will surface them anyway.

Risk · 01

Issuers are slow to adopt third-party underwriting inputs.

Mitigation

Position as feature-augmentation in the existing underwriting flow, not replacement. The issuer keeps the decision. Pulse expands the input set.

Risk · 02

Long-tail supplier data is fragmented and inconsistent.

Mitigation

The fragmentation is exactly what makes the integration work expensive enough to deter the next entrant. The cost compounds for the second mover, not the first.

Risk · 03

Default rates on AR-signal underwriting are unproven at scale.

Mitigation

Run a parallel underwriting program with an issuer partner first. Earn credit-loss credibility before committing to any volume guarantee.

Risk · 04

Top-five issuers (Chase, Amex, Capital One) build the capability in-house using their existing Plaid or Codat integrations.

Mitigation

Win the issuer-adjacent segment first. Embedded finance platforms, AP automation vendors, and supplier networks like Finexio originate card volume but do not underwrite. They are the buyer with the volume and without the in-house data-science team. Eighteen months of category proof there sets up either acquisition by a top-five issuer or a defensible network position the issuers route through. The in-house build by Chase is the exit, not the threat.

08 · Proof of motion

What I've already shipped on this thesis

The artifacts that turn this from an essay into something with traction. Published work, working-group seats, operator scars.

09 · Questions partners ask

The next three follow-ups

Pre-empted because the buyers and partners worth talking to will surface them anyway.

Why won't Plaid or Codat just add an underwriting score?

They might. Both sell the data feed today and could add scoring on top. The bet here is that an underwriting score that issuers will actually underwrite to requires more than a logistic regression on bank-statement data. It requires the AR/AP signal stack and the operator credibility to defend it under loss review. The data feed is necessary but not sufficient.

Why don't issuers just build this themselves?

Issuers want the answer, not the build. The Visa and Mastercard commercial-card programs are publicly committing to expand acceptance into the long tail. They do not want to expand their underwriting team to do it. A scored input that plugs into their existing flow is the path of least friction.

Bureaus already underwrite commercial cards. Why a separate layer, not just a Dun & Bradstreet or Experian feature?

Bureau files lag by 60 to 90 days and miss the long tail entirely. Dun & Bradstreet and Experian Business cover the established mid-market. The PYMNTS-Amex $1.6T trapped working capital figure sits below that line. A supplier invoicing 40 customers monthly through an ERP shows underwritable cash-conversion signal in week one. The bureau shows nothing for twelve months. Issuers do not need another bureau feature. They need a synthetic file for the segment bureaus structurally cannot serve.

Status

Pulse Underwriting is a published essay, not a stealth company. I am running Finexio. The thesis is here so the right operator or investor can find it and we can talk.

Of the eight ventures I've published, two are in discovery and I expect to operate one of them after Finexio. The rest, including this one, are pattern recognition I want in the open. If you read this and want to start it yourself, that is the outcome I'm hoping for.

Other concepts