Beta · rolling invites from June 2026

The mule account is the real weapon.
Pratyaksh is the receiving‑side defence.

Eight vendors compete to score the sender. None rate the receiver, freeze inside the golden window, or join bank‑side mule signals with payer identity. We do all three.

Join the beta How it works

The numbers we operate on

₹34,771 cr
India banking-fraud amount, FY24-25 (3.1× YoY). RBI Trend & Progress Report.
2.18%
CFCFRMS funds restored to victims. ₹167 cr of ₹52,969 cr reported (2021–Nov 2025).
23
Banks running RBI MuleHunter.AI by Dec 2025. Effectiveness data: not public.
24h
Golden window between report and freeze. Where Pratyaksh operates.
Section 01 · The structural gap

Trying to make customers un‑foolable
is an unwinnable war.

Every digital fraud terminates in a KYC‑verified, RBI‑regulated Indian bank account. The mule account is the real weapon. Nitesh Singhal · LinkedIn, 13 Apr 2026

Fraudsters no longer need an OTP. They socially‑engineer victims over days or weeks until the customer authenticates the transaction willingly. That single word — willingly — flips the legal question from did the bank fail? to did the customer fail?

The signals exist. Bureau, Signzy, Clari5, Wibmo, Hyperverge, IDfy, Effectiv, NPCI EFRM — eight vendors actively sell sender‑side detection to Indian banks. RBI's own MuleHunter.AI is free to 23 banks.

No incumbent operates on the receiving side, with consumer transparency, in the 24‑hour golden window where money becomes recoverable.

Where the effort lives today

Sender bank / PSP UPI / IMPS / NEFT / RTGS Receiving mule account cash‑out Sender‑side scoring Bureau, Signzy, IDfy, Hyperverge, Wibmo, Clari5, NPCI EFRM Receiving‑side trust layer RBI MuleHunter exists, bank-only. RTI on effectiveness: refused. PRATYAKSH Pre-routing risk score · CFCFRMS in 24h KYA × mule-flag join
Section 02 · How Pratyaksh works

Three checkpoints. Two milliseconds.
One audit trail.

Pratyaksh is a Rust SDK + service that plugs into your routing path. Before money moves, we score the receiving account on behavioural heuristics + bank‑side mule signals. If the transaction goes through and turns out to be flagged, we file the CFCFRMS report inside the 24‑hour golden window automatically. Every decision is anchored to a verifiable consent ledger.

Section 03 · The three capabilities

What we shipped to do this differently.

Capability 01

Pre-routing destination‑bank mule‑risk score

Before your routing engine commits to a corridor, Pratyaksh scores the receiving account on dormancy × sudden inflow, cashout velocity, and fan‑in from many distinct senders. Composes any/all of NPCI EFRM hash signals and (where opened) MuleHunter.AI emissions.

Capability 02

CFCFRMS auto‑reporting in the 24h golden window

The single biggest leverage point in the Indian fraud stack: most CFCFRMS reports miss the first‑hour reversal window. Pratyaksh pre‑files the structured report the moment a flagged transaction posts — turning the fastest possible report into a default rather than a procedural fight.

Capability 03

KYA × mule‑flag consent join

Joins your KYA identity on counterparties with bank‑side mule‑flag emissions where consented. The policy plane and the routing plane stay in one decision, anchored to a Kavach consent ledger. Future‑proofs the "we never settled to a flagged account" guarantee.

Section 04 · Built on what's verified

Numbers we didn't make up.

Pratyaksh's design rests on five RBI / Medianama / SOP-data points that we ground‑truthed against multiple sources before writing a line of code. Caveats and corrections from our research briefing.

3.1×
YoY growth in Indian banking fraud (FY23‑24 to FY24‑25). Partly retro per Mar 2023 SC judgment on 122 cases.
RBI T&P 2024-25
23
Banks live with MuleHunter.AI by Dec 2025. Canara, PNB, BoI, BoB, AU SFB named.
Medianama RTI
2.18%
CFCFRMS funds restored to victims. Detection works; recovery does not.
CFCFRMS, Nov 2025
62
Banks onboarded to I4C cyber‑crime portal. Improves report‑to‑freeze latency.
Medianama, Feb 2026
Section 05 · Who Pratyaksh is for

Built for the rails that need this most.

Beta cohort prioritises operators who can act on a receiving‑side score immediately. Larger commercial banks already have MuleHunter; we'll come to them when the integration story justifies it.

Small Finance Banks
Mandated fraud detection, fast vendor onboarding, MuleHunter adoption slower than commercial banks. Receiving‑side gap is acute.
12 wk cycle
Beta wave 1
Lending NBFCs
RBI Fraud Risk Management Directions 2024 mandate quarterly FMR‑3 reporting. API‑first, no empanelment needed.
8 wk cycle
Beta wave 1
Insurance underwriters
Per‑merchant mule‑risk scoring lifts loss‑ratios on digital‑payment claims. Requires data partnership + IRDAI sign‑off.
6–12 mo
Beta wave 2
Payment aggregators
High fraud exposure on UPI rails; most build in‑house or use Sardine/Bureau. Pratyaksh slots as KYA‑join layer.
3 mo
Wave 2
Scheduled commercial banks
Already have MuleHunter.AI free. Pratyaksh approaches as add‑on (CFCFRMS auto‑file + KYA join + audit dashboard) not replacement.
4 mo
Watch

Join the beta. Move first.

We're rolling beta invites from June 2026 to a small cohort of SFBs, NBFCs, and forward‑leaning underwriters. Early partners shape the integration surface and get founder access through the build.

If receiver‑side mule risk is in your remit and the 24‑hour golden window is a number you've measured (or wish you could), we want to hear from you.

✓ Thanks. We'll respond within 48 hours from support@devpilot.co.in.
We respond within 48 hours. No spam, no marketing list. Beta access for SFB / NBFC / insurer cohort first; commercial banks waitlisted.
All systems operational | Last verified | Beta · rolling Jun 2026 | Contact