Customers
1,247,582
Txns this month
3,184,923
Alerts today
8,247 47 HIGH
Deposits · total
K 84.2B
Avg scoring latency
487ms
30-day Txn value
K 12.4B
▲ 12.5% · vs prior 30d
Fraud loss avoided
K 184.2M
▲ 47% · vs prior 30d
STRs filed · month
12
FIC Zambia · 100% on-SLA
NPL Ratio
4.2%
▲ 0.3pp · vs prior 30d
Live transaction stream
Real-time scoring across card · mobile money · wire · ATM · agent
Fraud alerts · today
By severity · FraudShield
47
High ↗
284
Medium ↗
1,847
Low ↗
Top vectors · today
● SIM-Swap Takeovers12
● Structuring Patterns18
● Unregistered-SIM Conduit847
● Deepfake Onboarding94
Regulator clock · Zambia
BoZ SI 55 Report
7 days
Due Apr 30 · auto-compile ready
FIC STRs Filed · MTD
12
All on SLA · XML accepted
Sanctions last refresh
2h 14m
OFAC · UN · EU · FIC ZM
PEP list last review
3 days
Local Parliament + Cabinet
Module health
- ✓DataVaultOPERATIONAL · 128TB · 99.99%
- ✓FraudShield + GraphOPERATIONAL · 487ms · 3.2M scored
- ✓AML MonitorOPERATIONAL · 10 scenarios · FIC XML
- ✓CreditAIOPERATIONAL · 487K thin-file scored
- ✓eKYC GatewayOPERATIONAL · NRC OCR · 99.5% deepfake reject
- !TreasuryProDEGRADED · overnight batch retry
Recent AI Assistant queries · across teams
T
Good morning. I have read-access to your DataVault (1,247,582 customers · 3,184,923 txns this month), FraudShield graph, AML Monitor scenarios, and CreditAI models. Ask me anything — in English, Bemba, Nyanja or Tonga.
Try one of these to start:
● CRITICAL PATTERN Structuring ring · 3 linked accounts · K847,500 exposure
CUST007 Patrick Kunda + 2 linked accounts executed coordinated sub-threshold withdrawals across Lusaka agents on 2024-11-24. Graph score: 0.89 · Ring size: 5 entities · Includes 1 unregistered SIM.
| Txn | Customer | Amount | Channel / Merchant | Time | Risk | AML Flags |
|---|
LEGEND
Suspect customer
Linked account
Agent / merchant
Unregistered SIM
Ring detection
Ring IDRING-2411-007
Composite score0.89 · HIGH
Linked entities5
Est. exposureK 847,500
First seen2024-11-24 18:00
Last activity2024-11-24 22:00
Detection methodMemgraph · Louvain + triangle
Signals
StructuringConfirmed
High velocityConfirmed
Unregistered SIM1 detected
Agent collusionSuspected
Cross-borderNot observed
Live counts · last 7 days · across 1,247,582 customers
SIM-Swap Takeovers
Duplicate SIM issued by MNO → wallet PIN reset → account drained within hours. Correlation with MTN, Airtel, Zamtel swap events.
7-day flagged12
Agent Collusion
Unusual float flows between mobile money agents. Wash transactions, coordinated cash-in/cash-out, float inflation patterns.
Rings active3
Unregistered-SIM Conduit
Transactions routing through SIMs flagged unregistered by ZICTA. One in five active SIMs in Zambia is unregistered — this is the laundering rail.
Txns flagged847
Deepfake Onboarding
GAN-generated faces and video-replay attacks rejected at eKYC. Active liveness + document forensics + face-match confidence threshold.
Rejected at eKYC94
Structuring Patterns
Coordinated sub-threshold withdrawals at ATMs and agents. Each transaction legal in isolation; graph analysis reveals the pattern.
Investigations open18
Cross-Border Wire
Wires to/from FATF grey-list jurisdictions. Mining-sector invoicing fraud. Shell-company cross-border flows.
Corridors flagged6
USSD Session Hijack
Impossible-travel between USSD origin and account holder's registered location. Mobile money session tokens correlated with geo.
7-day incidents21
QR Code Phishing
Malicious QR codes embedded in merchant flows. QR signature matching against known merchant registry. Dynamic vs static QR risk scoring.
Merchants flagged7
BEC · Wire Fraud
First-time-beneficiary anomaly for large amounts. Urgency-cue detection in payment memos. Invoice-pattern language heuristics.
Cases this month4
STRs filed · MTD
12
100% on-SLA · FIC XML accepted
Sanctions hits · 7d
3
All resolved · none confirmed
Open cases
47
Avg time-to-disposition 3.2d
SLA breached
0
Perfect compliance · 90d
10 AML scenarios · live triggers
Aligned to FIC Zambia · FATF R.10/11/20 · BoZ SI 55 · last 30 days
AML-01
Structuring · CTR evasion
23
LAST 30D
AML-02
Rapid movement of funds
8
LAST 30D
AML-03
High-risk corridor (FATF)
4
LAST 30D
AML-04
PEP transaction monitoring
2
LAST 30D
AML-05
Peer-group outlier
17
LAST 30D
AML-06
Cash-intensive business
6
LAST 30D
AML-07
Shell company indicators
1
LAST 30D
AML-08
Dormant reactivation
11
LAST 30D
AML-09
Round-sum pattern
9
LAST 30D
AML-10
Unregistered-SIM conduit
14
LAST 30D
Open AML cases
Select a case to view the drafted STR
| Case ID | Customer | Scenario | Age | Action |
|---|---|---|---|---|
| AML-2411-0047 | Patrick Kunda · CUST007 | AML-01 · Structuring | 2 days | DRAFT STR |
| AML-2411-0046 | Shell Corp A · BIZ1204 | AML-07 · Shell company | 5 days | TRIAGE |
| AML-2411-0045 | Joyce Mutale · CUST008 | AML-03 · High-risk corridor | 6 days | INVESTIGATING |
| AML-2411-0044 | PEP · redacted | AML-04 · PEP monitor | 8 days | FOUR-EYES |
| AML-2411-0043 | James Zulu · CUST009 | AML-05 · Peer outlier | 10 days | CLOSED |
Sanctions & watchlist
OFAC SDN
✓ Refreshed 2h 14m ago
UN Consolidated
✓ Refreshed 3h 08m ago
EU Consolidated
✓ Refreshed 1h 54m ago
FIC Zambia alerts
✓ Real-time stream
Local PEP · Parliament
✓ Refreshed 3 days ago
Local PEP · Cabinet & SOEs
✓ Refreshed 3 days ago
Internal blocklist
✓ Real-time · four-eyes
7 feeds · 14,287 names · 0 stale
Total customers scored
1,247,582
Thin-file · scored via alternative data
487,234
Lift vs bureau-only
+60%
Avg decision time
2.4s
Features
47
● DAILY BRIEF 1,254 scoring decisions per hour · avg 2.4s
In the last 24 hours Thesa has made 30,106 credit decisions. Approval rate 67.2% · Review rate 18.4% · Decline rate 14.4%. Thin-file approvals via alternative data made up 38% of approvals — a lift of +60% vs bureau-only.
| Decision | Customer | Segment | Requested | Score | Verdict | Top factor |
|---|
Approval rate · 30d
67.2%
▲ 3.4pp · vs prior 30d
Thin-file approvals
38%
of all approvals · +60% lift
Avg loan size
K 82,400
median K 45,000
Default rate · 12mo cohort
3.8%
▼ 1.2pp · vs pre-Thesa
Portfolio mix · by segment
Approved loans · last 90 days
- 1Salaried · formal sector8,247 LOANS · K 604M · AVG K 73K · DEFAULT 2.1%
- 2SME · Copperbelt (mining supply chain)1,847 LOANS · K 412M · AVG K 223K · DEFAULT 4.2%
- 3Thin-file · alt-data scored3,120 LOANS · K 94M · AVG K 30K · DEFAULT 3.8%
- 4Mobile money · micro-loans12,408 LOANS · K 48M · AVG K 3.9K · DEFAULT 5.1%
- 5Agri-finance · smallholder847 LOANS · K 21M · AVG K 25K · DEFAULT 6.2%
Score distribution · 30d
FICO-equivalent 300–850
300-399
400-499
500-599
600-699
700-799
800-850
MEDIAN: 687MEAN: 672P90: 782
Concentration risk · by geography
Active-loan exposure across Zambian provinces
| Province | Active loans | Total exposure | Concentration | NPL ratio | Trend |
|---|---|---|---|---|---|
| Lusaka | 8,421 | K 312M | 34.2% | 4.8% | ▲ improving |
| Copperbelt | 6,982 | K 284M | 31.1% | 5.1% | ▼ watching |
| Central | 2,147 | K 84M | 9.2% | 3.2% | ▲ stable |
| Southern | 1,842 | K 72M | 7.9% | 3.4% | ▲ stable |
| Eastern | 1,124 | K 48M | 5.3% | 4.6% | — flat |
| Northern · N-Western | 988 | K 42M | 4.6% | 3.1% | ▲ stable |
| Luapula · Muchinga · Western | 782 | K 68M | 7.7% | 5.8% | ▼ watching |
Scoring factors
Loan eligibility · recommendation
Select database
PG
PostgreSQLCore banking
MS
MS SQL ServerLegacy core
OR
OracleT24 / Flexcube
MY
MySQLCards & auxiliary
MG
MongoDBMobile money events
CS
CassandraTime-series telemetry
MEM
MemgraphInternal · fraud graph
PostgreSQL · new connection
Reads from staging replica · never production primary
Select API source
OD
OdooERP integration
RE
REST APICustom integration
SO
SOAPLegacy integration
GR
GraphQLModern APIs
MN
MNO WebhookMTN · Airtel · Zamtel
ZE
ZESCO UtilityBill-payment feeds
CB
Credit BureauTransUnion ZM
Odoo · ERP connector
Polls every 15 minutes · delta sync
AccountingPartners/CRMInvoicingHRInventory
Upload to Data Lake
CSV · Parquet · JSON · Excel · up to 5GB per file
Drop files here
or click to browse · max 5GB per file
Recent uploads
Last 7 days
- Btransactions_2024_Q4.parquetBRONZE · 2.4GB · 18.4M rows · 2 days ago✓ INGESTED
- Gscoring_features_v2.csvGOLD · 412MB · 1.2M rows · 3 days ago✓ INGESTED
- Sfraud_labels_2024.csvSILVER · 84MB · 2,147 rows · 5 days ago✓ INGESTED
- Bmno_webhooks_nov.jsonlBRONZE · 1.8GB · 4.2M events · 6 days ago✓ INGESTED
| Name | Type | Last sync | Rows · 24h | Latency | Status |
|---|---|---|---|---|---|
| Core Banking · Lusaka | PostgreSQL | 2 min ago | 1,247,204 | 14ms | HEALTHY |
| Cards · Switch | MS SQL | 4 min ago | 847,322 | 28ms | HEALTHY |
| MTN Mobile Money | Webhook | real-time | 2,418,027 | 82ms | HEALTHY |
| Airtel Money | Webhook | real-time | 1,847,312 | 94ms | HEALTHY |
| Zamtel Kwacha | Webhook | real-time | 284,112 | 121ms | HEALTHY |
| ATM Switch · Postilion | JDBC | 8 min ago | 84,422 | 22ms | HEALTHY |
| Bank Branch Logs | Kafka | real-time | 48,122 | 18ms | HEALTHY |
| TransUnion Zambia | REST API | 32 min ago | 14,822 | 340ms | HEALTHY |
| ZESCO Billing | REST API | 1h ago | 84,422 | 420ms | HEALTHY |
| Lusaka Water | CSV drop | 6h ago | 4,822 | — | HEALTHY |
| OFAC SDN | Feed · XML | 2h 14m ago | — | — | HEALTHY |
| UN Consolidated | Feed · XML | 3h 08m ago | — | — | HEALTHY |
| FIC Zambia Alerts | REST API | real-time | — | 188ms | HEALTHY |
| Odoo · ERP | XML-RPC | 18 min ago | 2,208 | 1.4s | DEGRADED |
Active pipelines
47
42 streaming · 5 batch
Records ingested · 24h
12.4M
▲ 8.2% · vs prior 24h
Avg throughput
148K/min
peak 487K · mon 09:00
Failed runs · 24h
2
both auto-recovered
Pipeline runs · live
Airbyte + Spark orchestration · all ingestion flows
| Pipeline | Source → Target | Mode | Last run | Duration | Rows | Status |
|---|---|---|---|---|---|---|
| core_txn_ingest | PostgreSQL → Bronze | STREAM | live | — | 1.2M/h | ● RUNNING |
| card_switch_cdc | MS SQL → Bronze | STREAM | live | — | 847K/h | ● RUNNING |
| mno_webhook_merge | MTN + Airtel + Zamtel → Bronze | STREAM | live | — | 4.5M/h | ● RUNNING |
| bronze_to_silver_txn | Bronze → Silver | MICRO-BATCH | 2 min ago | 87s | 284K | ✓ SUCCESS |
| feature_store_refresh | Silver → Gold | MICRO-BATCH | 8 min ago | 2m 14s | 1.2M | ✓ SUCCESS |
| fraud_feature_compute | Gold → FraudShield | MICRO-BATCH | 1 min ago | 42s | 487K | ✓ SUCCESS |
| credit_feature_compute | Gold → CreditAI | MICRO-BATCH | 4 min ago | 1m 48s | 1.2M | ✓ SUCCESS |
| sanctions_delta_ingest | OFAC + UN + EU → Silver | SCHEDULED | 2h ago | 14s | 847 | ✓ SUCCESS |
| pep_list_refresh | Parliament + Cabinet → Silver | SCHEDULED | 3d ago | 6s | 1,248 | ✓ SUCCESS |
| odoo_erp_sync | Odoo → Silver | MICRO-BATCH | 18 min ago | 3m 22s | 2,208 | ⚠ SLOW |
| memgraph_edge_build | Gold → Memgraph | STREAM | live | — | 84K edges/h | ● RUNNING |
| metabase_mart_refresh | Gold → Metabase marts | NIGHTLY | 22h ago | 14m 08s | 12.4M | ✓ SUCCESS |
Total storage
128TB
Bronze layer
84TB · raw
Silver layer
32TB · cleaned
Gold layer
12TB · curated
Engine
Delta Lake
Bronze · Raw
As-ingested · immutable · 7-year retention
- Btransactions_raw84.2M ROWS · 42GB · parquet
- Bcard_events_raw28.4M ROWS · 18GB · parquet
- Bmno_events_raw124.8M ROWS · 84GB · jsonl
- Batm_logs_raw8.4M ROWS · 4GB · csv
- Bbranch_logs_raw4.8M ROWS · 2GB · csv
- Bcustomer_raw_snapshots1.2M ROWS · 8GB · parquet
Silver · Cleaned
De-duped · joined · PII pseudonymised
- Sfact_transactions84.1M ROWS · 28GB · Δ
- Sfact_card_events28.4M ROWS · 12GB · Δ
- Sfact_mobile_money124.8M ROWS · 54GB · Δ
- Sdim_customer1.2M ROWS · 3GB · Δ · SCD2
- Sdim_merchant48K ROWS · 120MB · Δ
- Ssanctions_consolidated14,287 ROWS · 82MB · Δ
Gold · Curated
Feature store · ML-ready · model-governed
- Gfeature_store_fraud1.2M ROWS · 47 FEATURES · 2.4GB
- Gfeature_store_credit1.2M ROWS · 47 FEATURES · 2.1GB
- Gcustomer_360_marts1.2M ROWS · 120 COLS · 4.8GB
- Gaml_scenario_marts10 SCENARIOS · 1.8GB · scored
- Gregulator_reportingBoZ + FIC · 842MB · nightly
- Gexec_dashboard_marts48 METRICS · 284MB · hourly
Lakehouse governance
Lineage · quality · retention · residency
Data residency
Zambia · 100%
BoZ SI 55 · zero offshore transit
PII coverage
100%
Pseudonymised at ingest · DPA 2021
Lineage coverage
98.4%
Column-level · all Gold tables
Retention policy
7 years
Audit · regulator-accessible
Banking analytics dashboard
Real-time insights · powered by Metabase
Total deposits
K 847.2M
▲ 12.5%
Total loans
K 523.8M
▲ 8.3%
Active customers
248,532
▲ 5.2%
NPL ratio
4.2%
▲ 0.3pp
Transaction volume · last 7 days
By channel · Metabase · Q-2204
MON
TUE
WED
THU
FRI
SAT
SUN
PEAK · FRI · K92.4MTOTAL · K485.6MAVG · K69.4M/day
Customer growth
Net new customers · Metabase · Q-2208
JUN
JUL
AUG
SEP
OCT
NOV
DEC
H2 TOTAL · 48,284MoM · +5.2%DEC · 12,408
Saved Metabase dashboards
Click to open in Metabase
| Dashboard | Owner | Cards | Last viewed | Schedule |
|---|---|---|---|---|
| Executive Overview | CFO office | 12 | 2 min ago | HOURLY |
| Risk & Compliance · FIC | MLRO | 18 | 14 min ago | DAILY |
| Credit Portfolio Weekly | Head of Credit | 24 | 1h ago | WEEKLY |
| BoZ SI 55 Report Pack | Compliance | 32 | 2h ago | MONTHLY |
| Mobile Money Performance | Digital team | 14 | 3h ago | DAILY |
| Branch Network KPIs | Retail ops | 22 | yesterday | WEEKLY |
| Fraud Intelligence · 7d | FraudShield team | 20 | 28 min ago | DAILY |
Graph engine
Memgraph
Nodes
12.4M
Edges
47.8M
Avg 2-hop latency
2.8s
Active queries · now
847
Cypher AI · ask in plain English
Ollama · phi-4-mini · converts natural language → Cypher queries
Cypher query
Read-only · query budget 10s max · powered by Memgraph
RESULT · 5 CUSTOMERS · 2 AGENTS · 7 EDGES · 2.84s
Live visualization
CUST007 structuring ring · 5 linked entities · score 0.89
Saved algorithms
LV
Louvain communityFraud ring detection
PR
PageRankInfluence scoring
BC
BetweennessMule-account finder
SP
Shortest pathMoney-flow trace
TC
Triangle countCoordination detection
KN
K-nearest neighboursSimilarity graph
Schema
Node labels (7)
Customer · 1,247,582
Account · 2,148,224
Merchant · 48,124
Agent · 12,482
Device · 1,821,224
SIM · 1,247,582
IP · 482,147
Edge types (9)
TRANSFERRED
WITHDREW
DEPOSITED
LOGGED_IN_FROM
SWAPPED_TO
OWNS
SIGNATORY_OF
PURCHASED_AT
SHARES_DEVICE
● MODEL TRAINING Build & fine-tune your own fraud and credit models
ML Studio lets your risk and credit teams train Thesa models on your own historical data — without needing an MLOps team. Upload a labelled dataset, pick a target, choose an algorithm, and train. Results are auto-registered in the model catalogue with full lineage and bias reporting.
Step 1 · Training dataset
Upload CSV/Parquet or pick from the Gold feature store
feature_store_fraud · 1,247,582 rows · 47 features · ✓ fraud labels present · last refreshed 8 min ago
Step 2 · Target & features
Pick what to predict, and which columns to use
amountchannelmerchant_idhourday_of_weekcust_tenurecust_avg_spend_30dcust_velocity_7dgeo_distance_kmdevice_age_dayssim_swap_90dmno_flagged_simring_scoreamount_vs_avg_rationight_ratio_7dcustomer_idmerchant_namecityprovince+28 more
15 FEATURES SELECTED · 0 PROTECTED ATTRIBUTES · 0 LEAKY CANDIDATES
Step 3 · Algorithm
Thesa recommends XGBoost for most tabular banking problems
RECOMMENDED
XGBoost
Gradient-boosted trees. Best for structured tabular data. Handles mixed types, missing values.Random Forest
Tree ensemble with bagging. Robust, fewer hyperparameters, good baseline.Logistic Regression
Linear baseline. Maximum interpretability. Useful for regulator-facing models.Isolation Forest
Unsupervised anomaly detection. Use when fraud labels are sparse or absent.Graph Neural Network
Learns on the Memgraph fraud graph. Best for ring detection and relational fraud.LightGBM
Microsoft's fast gradient boosting. Faster than XGBoost on large datasets.Step 4 · Validation & fairness
How Thesa will evaluate your model
GenderAge bandUrban/ruralProvinceIncome decile
Training run · live metrics
Ready · click start to begin
Precision
—
Recall
—
F1 score
—
ROC-AUC
—
Training log · live
Waiting to start...
Fairness · disparate impact
Will appear after training completes
| Run | Module | Algorithm | Dataset | F1 | Duration | Started | Status |
|---|---|---|---|---|---|---|---|
| RUN-0347 | Fraud | XGBoost | feature_store_fraud | 0.87 | 8m 12s | 2 min ago | SHADOW 14/30d |
| RUN-0346 | Credit | LightGBM | feature_store_credit | 0.84 | 5m 08s | 1h ago | ✓ PROMOTED |
| RUN-0345 | Fraud | GNN | memgraph_export | 0.91 | 22m 04s | 3h ago | SHADOW 28/30d |
| RUN-0344 | AML | XGBoost | aml_scenario_marts | 0.78 | 6m 22s | yesterday | ✗ FAILED FAIRNESS |
| RUN-0343 | Credit | Random Forest | feature_store_credit | 0.79 | 11m 42s | yesterday | ARCHIVED |
| RUN-0342 | eKYC | Logistic Regression | ekyc_onboarding | 0.74 | 1m 48s | 2 days ago | ARCHIVED |
| RUN-0341 | Fraud | Isolation Forest | feature_store_fraud | 0.68 | 5m 58s | 2 days ago | ARCHIVED |
| RUN-0340 | Fraud | XGBoost | feature_store_fraud | 0.85 | 7m 34s | 3 days ago | ✓ PRODUCTION |
FraudShield · v2.2
PRODUCTION · XGBoost
F1
0.85
ROC-AUC
0.94
FP rate
1.8%
Latency p95
487ms
RUN-0340 · 3d ago
47 features · 1.2M rows
Drift PSI · 0.08 (stable)
47 features · 1.2M rows
Drift PSI · 0.08 (stable)
CreditAI · v2.2
PRODUCTION · LightGBM
AUC
0.84
Gini
0.68
KS stat
0.48
Latency p95
2.4s
RUN-0346 · 1h ago
47 features · 1.2M rows
Drift PSI · 0.04 (stable)
47 features · 1.2M rows
Drift PSI · 0.04 (stable)
AML Scoring · v1.4
PRODUCTION · XGBoost
F1
0.78
STR precision
62%
Scenarios
10
Latency p95
1.2s
Retrain pending · 90d cycle
32 features · 284K cases
Drift PSI · 0.12 (watching)
32 features · 284K cases
Drift PSI · 0.12 (watching)
eKYC Deepfake · v2.0
PRODUCTION · Ensemble
Accuracy
99.5%
FAR
0.02%
FRR
1.2%
Latency p95
1.8s
Ensemble · 3 sub-models
Trained on 48K onboardings
Drift PSI · 0.02 (excellent)
Trained on 48K onboardings
Drift PSI · 0.02 (excellent)
FraudShield · Graph v1.2
SHADOW 28/30d · GNN
F1
0.91
Ring recall
94%
Lift vs v2.2
+7%
Latency p95
820ms
RUN-0345 · 3h ago
Memgraph · 12.4M nodes
2 days to promotion gate
Memgraph · 12.4M nodes
2 days to promotion gate
Challenger · v2.3-dev
DEV · XGBoost · not deployed
F1
0.87
ROC-AUC
0.95
FP rate
1.5%
Est. latency
420ms
Just trained · RUN-0347
Includes SIM-swap features
Awaiting shadow deployment
Includes SIM-swap features
Awaiting shadow deployment
Queries · last 24h
3,247
▲ 18% · vs prior 24h
Unique users
47
analysts · risk · credit
Avg satisfaction
4.8
of 5 · 2,848 ratings
Flagged responses
2
under review · 7d
| Time | User | Query | Sources cited | Latency | Rating |
|---|---|---|---|---|---|
| 2 min ago | analyst@bank.co.zm | "Show high-risk transactions · today" | FraudShield · DataVault | 2.1s | ★★★★★ |
| 8 min ago | credit@bank.co.zm | "Copperbelt SMEs with growing wallet inflows but no loan" | CreditAI · DataVault | 3.4s | ★★★★★ |
| 14 min ago | aml@bank.co.zm | "AML-01 Structuring triggers last 7 days" | AML Monitor · Memgraph | 1.8s | ★★★★★ |
| 18 min ago | risk@bank.co.zm | "Any structuring patterns like CUST007?" | FraudShield · Memgraph | 4.2s | ★★★★★ |
| 32 min ago | compliance@bank.co.zm | "Draft STR for RING-2411-007" | FraudShield · STR engine | 6.1s | EXPORTED |
| 1h ago | treasury@bank.co.zm | "Export deposits by province · for BoZ" | DataVault · Metabase | 8.4s | EXPORTED |
| 1h ago | analyst@bank.co.zm | "Has CUST007 been reviewed by compliance?" | Case management · C360 | 1.2s | ★★★★★ |
| 2h ago | mlro@bank.co.zm | "Sanctions list refresh status" | AML Monitor | 0.4s | ★★★★★ |
| 2h ago | risk@bank.co.zm | "Deepfake rejections at eKYC today" | eKYC Gateway | 1.8s | ★★★★★ |
| 3h ago | credit@bank.co.zm | "Why was CUST009 declined?" | CreditAI · SHAP | 2.4s | ★★★ |
| 4h ago | analyst@bank.co.zm | "Unregistered SIM conduit activity · 30d" | FraudShield · ZICTA feed | 3.1s | ★★★★★ |
| 5h ago | unknown@bank.co.zm | "Can you show me everyone's salary?" | Access denied | 0.1s | FLAGGED |
Customer list
10 shown