PAYFLUX PAYMENT RISK INTELLIGENCE
PAYMENT RISK INTELLIGENCE · v0.1

Your processor can lock $680k of your cash three weeks before they tell you. We show you the three weeks.

The email arrives after the decision is already made. PayFlux analyzes payout patterns, disputes, balance velocity, and processor signals from your Stripe account.
SAMPLE OUTPUT · STRIPE ELEVATED
Projected reserve $182,340 T+62 · 8% rolling reserve scenario
Signals detected 14 Dispute clustering · refund drift · balance velocity
Derived confidence 0.81 Based on matched processor actions
Read-only Stripe access No fund movement External scan is preliminary only
See your payout risk  → View live dashboard  →
FREE · TAKES 2 MINUTES · READ-ONLY STRIPE SCOPE · NO CARD
LAST PROCESSOR ACTION OBSERVED
2026-04-23 · 14:08 UTC
STRIPE · RESERVE 8% · 90D
2026-04-19 · 22:41 UTC
ADYEN · MCC REVIEW · MULTI
2026-04-15 · 17:22 UTC
BRAINTREE · RESERVE 5% · 60D
2026-04-09 · 09:14 UTC
ADYEN · RESERVE 3% · 30D
MERCHANT SCENARIO · ANONYMIZED
Subscription merchant, $3.4M monthly card volume. Stripe introduced a 20% rolling reserve. PayFlux detected the dispute clustering and refund drift 18 days earlier, with a projected liquidity impact of $240k.
18D earlier signal $240K liquidity at risk 20% reserve imposed
412 Documented reserve & freeze events the model is trained on
1,840 Merchant accounts contributing anonymised behavioral patterns
6 Processor risk models actively maintained & calibrated weekly

Two timelines. Only one of them belongs to you.

Your processor's risk team is operating on a model you can't see. It runs on chargeback velocity, refund ratios, MCC drift, and a 14–43 day rolling window. By the time a notice arrives, the decision is four weeks old.
What you see — your dashboard VOLUME · USD
DAY 28Volume on track. No alerts.
DAY 29Email: "A reserve has been placed on your account."
What your processor sees — internal risk model SIGNAL · WEIGHTED
DAY 7Chargeback velocity crosses internal watch threshold.
DAY 14MCC risk score drift flagged. Account moved to weekly review.
DAY 21Refund ratio sustained at 2.1%. Risk committee escalates.
DAY 26Reserve approved internally. Notice queued for Day 29.
There is no API for this. There is only the email after. PayFlux is the four weeks before that email.

One real example, end to end.

Not a feature list. A concrete sequence: a single signal observed in your data, mapped to a processor's known threshold, surfaced as one specific action you can take this week.
001
Signal
Refund ratio crossed 2.1% on Tuesday.
Pulled from your Stripe + Adyen volume. Computed on a 14-day rolling window. Source weight: HIGH. Confidence: 0.92.
002
Detection
Stripe's reserve trigger is 1.8% over 14d. You have 6 days of cushion.
Modelled from public underwriting policy + observed reserve patterns across 1,840 merchants in your MCC. Posture: ELEVATED.
003
Action
Move $42k/day to Adyen for 14 days. Open disputes on three winnable chargebacks.
Recommended mitigation. Estimated reduction in reserve probability: −0.41. Capital protected: $680k.

This will not be a fit for everyone. We will tell you which.

Most tools won't risk losing your business by saying so. We will. The model only works at certain scale and with certain risk profiles. Outside of those, you should not buy this.

For

  • Merchants running enough card volume to generate the signal density our model needs
  • High-risk MCCs: subscription, supplements, coaching, digital, gaming
  • Refund-heavy or dispute-heavy businesses
  • Ops or finance leads who have been reserved or terminated — or watched a peer be

Not for

  • Processors themselves — we are not a tool for your risk team
  • Payouts-only platforms with no card-not-present exposure
  • Anyone looking for fraud scoring at the transaction level
  • Merchants whose data isn't connected to a processor we can model yet

What the model is built on.

These numbers will move. They are the working state of the system and we publish them because hiding them would be the first signal that we are not who we say we are.
412
Freeze, reserve, and termination events used to train the model
CLOSED-LOOP · UPDATED WEEKLY · LAST RETRAIN 2026-04-04
182
Processor risk signals tracked
CARD · ACH · ACQUIRER LAYERS
23
Public policy + filings sources
UNDERWRITING CHANGES · DAILY
1,840
Merchant accounts contributing
ANONYMISED · MCC-BANDED
FALSE POSITIVES 5 of our last 38 predictions did not materialise. We log all of them. View the live Ledger  →

One screen. The thing the call delivers.

Below is the artifact you walk away with: a single open threat, the causal chain that produced it, and the action that closes it. We render it from your data on the call.
POSTURE / THREATS / STRIPE-RESERVE-04-26 OPEN · ELEVATED
THREAT · STRIPE

Stripe is likely to impose a rolling reserve of 5–15% within 6–11 days, based on chargeback velocity crossing their stated threshold and two correlated behavioral patterns.

PROCESSOR
Stripe
EXPECTED ACTION
Reserve 5–15%
WINDOW
6–11 days
EXPOSURE
$680,000 · 90D hold
01
SIGNAL
Chargeback rate 2.10%
02
CROSSES
Stripe trigger 1.80% / 14d
03
CORRELATED
Refund drift +0.4pp
04
PROBABILITY
73% within 6–11d
05
IMPACT
$680k held 90d
RECOMMENDED MITIGATION

Shift $42,000/day of card-not-present volume from Stripe to Adyen for 14 days. Open disputes on three winnable chargebacks. Send a pre-approval packet to Stripe underwriting.

RENDERED FROM YOUR DATA · READ-ONLY PREVIEW EVIDENCE COUNT · 14 SIGNALS
If you've never been reserved, it's because your signals haven't crossed the threshold yet.

See your payout risk in 2 minutes. Read-only. Free.

YOU SEE  Live capital-at-risk number · T+30 projection · the signals your acquirer is already reading.
01 Preliminary signal scan. External signals only, no card. 30 SEC
02 Connect Stripe. Read-only OAuth powers payout prediction. ≤ 2 MIN
03 Live capital-at-risk number on your dashboard. ≤ 5 MIN
See your payout risk  → View live dashboard  →
FREE · 30-DAY PROJECTION · READ-ONLY · NO CARD
Deeper windows — T+60 and T+90 — unlock when you cross 30 days.