Score every payment before it settles
Transaction stream
Live
48.2k
Evaluated today
214
In review
0.4%
Flagged
Latest evaluations
$84,200 · UPI
Rejected
$12,500 · Card
Review
$6,300 · UPI
AML review
$1,240 · Card
Cleared
Zoven checks each transaction against your rules, scores it for fraud and money-laundering risk, and routes the risky ones to your team while the clean ones pass through..
Trusted by risk and compliance teams
Block the bad, clear the good, review the rest
Most payments are fine and should just go through. Zoven does the checking so your team only sees the ones that genuinely need a look.
Rules that fit your policy
Build the checks you want from real conditions: amounts, velocity, geography, card and account patterns. Each rule blocks, reviews, or flags.
3
rule types, one engine
Fraud caught in the browser
Bot scoring, keystroke behaviour, and checkout tampering catch stolen cards and automated attacks the amount alone would miss.
Bot 0–100
scored per checkout session
AML through to filing
Suspicious payments route into an AML queue, and a confirmed case turns into a ready-to-file FIU report package.
FIU-ready
STR files, generated for you
Every payment, checked and sorted
A transaction comes in through the API, gets enriched, and runs against your rules. Zoven scores it, applies the strongest verdict, and either clears it, sends it for review, or rejects it. The rules it hit and why stay on the record.
TRANSACTION
Blocked
₹
₹84,200
UPI → Bluepeak Retail
88
risk
RULES FIRED
3 matched
Velocity · 6 in 10 min
BLOCK
Sender geo mismatch
+14
Card BIN high-risk
+12
Browser bot score 72
+9
Strongest verdict wins
Blocked before settle
Weighted 0–100 risk score
Factors like velocity, geography, card BIN, and browser signals combine into one score per payment.
Three rule types, one verdict
Logical expressions, mule checks, and risk scores run together, and the strongest action wins.
Clear, review, or reject
Clean payments pass through, risky ones route to a queue, and blocked ones stop, all automatically.
Every rule hit on the record
The transaction shows which rules ran, which matched, and the factors behind the score.
CHECKOUT SESSION · REPLAY
Automated
KEYSTROKE TIMELINE
Pasted in one action
No human typing rhythm
Mouse path is a straight line
Checkout in an injected iframe
Browser fingerprint emulated
Automated checkout — likely stolen card
Bot 91
The amount is only half the story
A legitimate-looking payment can still be a bot with a stolen card. Zoven reads the browser session behind each checkout, so the signals a rule can't see get caught too.
Bot scoring
Each session scored 0 to 100 for automated versus human behaviour
Keystroke behaviour
Typing, pasting, and correction patterns tell a real person from a script.
Checkout tampering
Origin, referrer, and injected checkout code checked for redirect fraud.
Named scenarios
Stolen card, stolen API keys, and redirect fraud each get a clear verdict.
From payment to decision, in three steps
Every transaction runs the same path. Zoven handles the checking so your team only sees the ones that need a look.
Ingest the payment
Send each transaction through the API, with a browser SDK reading the checkout session alongside it.
01
API and browser SDK
Enrich and evaluate
Zoven adds IP and card intelligence, then runs the payment against your rules and fraud signals.
02
scored against every active rule
Clear, review, or block
Clean payments pass through, risky ones open a case, and blocked ones stop before they settle.
03
the strongest verdict wins
A platform built for a monitoring desk
The parts that make Transaction Monitoring fit how your team already works, not the other way round.
AML & STR filing
A payment that trips your AML rules routes into an AML queue, and a confirmed case turns into an FIU-ready STR package your team can file, not assemble by hand.
Routes suspicious payments to an AML queue
Generates the KC2, TS3, and GS1 report files
Ready to upload straight to FIU-IND
AML case
₹84,200 · UPI · confirmed
Suspicious
STR PACKAGE · FIU-IND
3 files
KC2
Account & customer profile
✓
TS3
Transaction detail
✓
GS1
Ground for suspicion
✓
Package ready
Upload to FIU-IND →
Score thresholds
Set two thresholds on the risk score. One decides when a payment goes to review, the other when it is blocked outright.
RISK SCORE
0–100
Clear
0–39
Pass through
Review
40–69
Monitoring queue
Block
70–100
Reject before settle
Decided before it settles
Every payment runs an async ingest, enrich, and evaluate pipeline in milliseconds, so the verdict lands before the money moves.
PER PAYMENT
42 ms
✓
Ingest
1 ms
✓
Enrich · IP + card + BIN
18 ms
✓
Evaluate · 14 rules
23 ms
Verdict before it settles
Blocked
Rules you can version and diff
Rules are versioned like code and export as CSV or pseudocode, so your team can review a change, diff it, and roll it back.
RULES
v14 · 2h ago
+
velocity_card_10m
added
~
geo_highrisk
edited
–
legacy_amount_gt
removed
Diff, roll back, restore
Export ↓
A rules engine that speaks your risk language
Build the conditions your policy needs and nest them with AND and OR. From a simple amount check to a time-windowed velocity rule, it runs on every payment.
Comparisons
Amount, count, or field against a value: greater than, between, equals.
List checks
In list, not in list, or any-in for blocklists and allowlists.
String patterns
Contains, domain matches, and consecutive-digit detection.
High-risk geography
Flag transactions touching high-risk countries.
Velocity windows
Aggregate over minutes, hours, or days by card, email, phone, or IP.
Field-to-field & math
Compare one field to another, or run a value through a math pipeline first.
Mule identifiers
Risk-score thresholds
Two thresholds set when a score reviews and when it blocks.
Rules are versioned, and can be imported and exported as CSV or pseudocode.
Depth a simple limit check can't reach
The parts that make the difference on real fraud and money-laundering cases, not just the obvious ones.
Aggregate by card, issuer country, email, phone, sender, receiver, or IP over any time window
Compare one field to another, or run it through a math step, before the condition is even checked
Connect it to how you already work
Send transactions and manage rules through the channels your systems and team already use.
REST API
Verify a transaction and get its evaluation status back.
Browser SDK
Drop into your checkout to read the session behind each payment.
API keys
Secret keys for your servers, publishable keys for the browser.
Rules import & export
Move rules in and out as CSV or pseudocode.
IP & card intelligence
Geolocation and card BIN data added to every transaction.
Browser & device signals
Origin, referrer, and behaviour captured at checkout.
Transaction history
Prior activity on the same card, account, or merchant.
Async pipeline
Ingest, enrich, and evaluate run as a queued flow.
Available data and channels vary by market and plan.
A payment is a decision made in an instant, and the wrong one is expensive both ways: fraud that settles, or good money you turned away.
Fraud looks like a normal payment
A stolen card and a real buyer fill in the same form. The difference is in the session behind the checkout, not the amount on the screen.
Block too hard and you lose good revenue
Every false decline is a real customer turned away, and they rarely come back. Tuning for safety without killing conversion is the hard part.
Static rules go stale
Money laundering hides in ordinary transactions, and the attack shifts faster than a fixed checklist. Yesterday rules miss today attempts.
Good to know
How does a transaction get evaluated?
A payment comes in through the API and is enriched, then run against your active rules: mule checks, logical expressions, and risk scores. Each rule returns an action, and the strongest one wins, so a block beats a review beats a pass. The result sets the transaction's status: cleared, under review, under AML review, or rejected.
What can a rule actually check?
How is fraud caught beyond the rules?
What happens on an AML flag?
Can our team review before a payment is rejected?
Bring a sample of real transactions to a walkthrough and watch them score, sort, and land in a queue.