110.5%
Liquidity Coverage Ratio
118.2%
Net Stable Funding Ratio
₹4,200 Cr
Total HQLA
6
AI models in production
Liquidity modules
Six connected modules, one liquidity intelligence engine. Click any tile to open it.
AI models powering this workspace
Liquidity Stress EngineBacktest 97.4%
Deposit Behavioral ModelMAE 3.6%
HQLA Optimization ModelCost reduction 6.8%
Intraday Liquidity PredictorShadow testing
AI Insights
Generated
A ₹500 Cr wholesale deposit renewal falls due in 30 days — the behavioral model flags a 34% chance of non-renewal at current pricing.
An additional 6.7% deposit run-off beyond baseline stress assumptions would breach the 100% regulatory LCR minimum.
Level 2B assets carry ₹18 Cr of avoidable haircut drag — the optimization model recommends a swap into Level 1 government securities.
110.5%
Liquidity Coverage Ratio
118.2%
Net Stable Funding Ratio
₹4,200 Cr
Total HQLA
₹3,800 Cr
30-day stressed outflows
LCR trend, trailing 6 months
NSFR — live gauge
available vs required stable
funding · RBI minimum 100%
funding · RBI minimum 100%
HQLA composition
Level 1 (govt securities)66%
Level 2A24%
Level 2B10%
Net cash flow gap, by maturity bucket
Maturity Gap Profile — Consolidated Book
As of 07-07-2026 · Balance sheet size ₹42,000 Cr · Total deposits ₹31,000 Cr
30-day cumulative gap: negative
0–30 day cumulative gap — contractual inflows less outflows
−₹340 Cr
Shortfall
31–90 day cumulative gap — contractual inflows less outflows
−₹180 Cr
Watch
Largest single-day outflow — bulk wholesale FD maturity
₹620 Cr
Watch
Wholesale funding maturing in 30 days
₹1,100 Cr
Watch
HQLA available to cover 30-day gap
₹4,200 Cr
Healthy
Retail deposit base — stable, granular funding
₹19,400 Cr
Healthy
AI Treasury Copilot
Deposit Behavioral Model · v3.4
Live
Recommended: Pre-fund ₹500 Cr via term repo, extend tenor on wholesale renewal 81% confidence
The 0–30 day bucket shows a ₹340 Cr shortfall driven by a single ₹620 Cr wholesale FD maturity with an elevated non-renewal probability. HQLA comfortably covers the gap today, but pre-funding now avoids drawing down Level 1 assets at short notice and keeps LCR well clear of the regulatory floor.
Why the model recommends this action — top signal contributions
Bulk FD maturity concentration
+0.28
Elevated non-renewal probability
+0.24
0–30 day negative cumulative gap
+0.20
Strong HQLA buffer
−0.15
Granular, stable retail base
−0.10
₹4,200 Cr
Total HQLA
₹6,000 Cr
Wholesale / uninsured deposits
100%
RBI regulatory LCR minimum
Deposit run-off stress simulator
Baseline 30-day stressed outflows of ₹3,800 Cr against ₹4,200 Cr HQLA. Move the slider to add extra run-off stress on the ₹6,000 Cr wholesale/uninsured deposit pool and see LCR recompute live.
0%40%
95.5%
Stressed LCR
28.6 days
Survival horizon
Recommended: Monitor closely — activate contingency funding plan triggers.
AI threshold: LCR falls below the 100% regulatory minimum at roughly 6.7% additional run-off — beyond that, CFP activation is required.
LCR vs deposit run-off stress
Stressed LCR as additional run-off increases. The dashed line marks the RBI 100% regulatory minimum.
■ stressed LCR - - 100% regulatory minimum ● current simulation
Estimated stressed LCR under this scenario, by funding scenario type
₹4,200 Cr
Total HQLA (haircut-adjusted)
₹18 Cr
Avoidable haircut drag
66%
Level 1 share of HQLA
HQLA composition and haircuts
AI-recommended collateral optimizations
| Action | From | To | Haircut saved | Priority |
|---|---|---|---|---|
| Swap into Level 1 government securities | Level 2B equity basket | Level 1 G-Sec | ₹9.4 Cr | High |
| Roll into shorter-tenor T-Bills | Level 2A corporate bonds | Level 1 T-Bills | ₹5.2 Cr | Medium |
| Increase repo-eligible collateral pool | Non-HQLA fixed assets | Level 2A eligible bonds | ₹3.4 Cr | Medium |
No rows match your search.
₹31,000 Cr
Total deposits
₹6,000 Cr
Wholesale / uninsured
18
Top-10 depositor concentration flags
Funding mix, by source
Largest depositor / counterparty concentrations
| Counterparty | Type | Exposure | % of total deposits | AI signal |
|---|---|---|---|---|
| DEP-1120 | Corporate treasury | ₹620 Cr | 2.0% | Watch |
| DEP-0847 | Mutual fund | ₹480 Cr | 1.5% | Watch |
| DEP-1305 | Insurance company | ₹410 Cr | 1.3% | Stable |
| DEP-0592 | Corporate treasury | ₹390 Cr | 1.3% | Stable |
No rows match your search.
365
Daily LCR calculations (YTD)
100%
Above regulatory minimum
0
Breaches this year
RBI LCR / NSFR compliance audit trail
| Timestamp | Metric | Basis | Result | Status |
|---|---|---|---|---|
| 07-07 06:00 | Daily LCR calculation | HQLA ₹4,200 Cr / stressed outflows ₹3,800 Cr | 110.5% | Compliant |
| 07-06 06:00 | Daily LCR calculation | HQLA ₹4,180 Cr / stressed outflows ₹3,850 Cr | 108.6% | Compliant |
| 07-01 09:00 | Monthly NSFR calculation | Available vs required stable funding | 118.2% | Compliant |
| 06-28 14:30 | Quarterly RBI liquidity stress test | Combined severe scenario, 30-day horizon | 79.2% | Reported — CFP reviewed |
No rows match your search.
5
Models in production
1
In shadow testing
95ms
Avg. inference latency
640K
Predictions served today
Registered models
| Model | Type | Status | Performance | Last trained | Explainability |
|---|---|---|---|---|---|
| Liquidity Stress Engine | Scenario factor model | Production | Backtest 97.4% | 3 days ago | Factor attribution |
| Deposit Behavioral Model | Survival / renewal model | Production | MAE 3.6% | 4 days ago | SHAP attribution |
| HQLA Optimization Model | Constrained optimization | Production | Cost reduction 6.8% | 5 days ago | Counterfactual explanation |
| Cash Flow Forecasting Model | Time-series (LSTM) | Production | MAPE 5.1% | 6 days ago | Attention weights |
| Funding Concentration Scorer | Herfindahl-based risk model | Production | AUC 0.84 | 7 days ago | Feature importance |
| Intraday Liquidity Predictor | Real-time time-series model | Shadow testing | MAE 4.9% | 1 day ago | Attention weights |
No rows match your search.
Global explainability — Deposit Behavioral Model
Mean absolute SHAP contribution across all deposit renewal predictions — what drives run-off risk overall.
Model lifecycle
01
Train
02
Validate
03
Shadow test
04
Promote
05
Monitor
Shadow models score live treasury data in parallel without serving recommendations, until they clear the production accuracy bar.
6
Connected systems
1.6M
Records synced today
0
Sync errors (24h)
Connected systems
| System | Type | Status | Last sync | Records today |
|---|---|---|---|---|
| Core Banking System | Deposits, loans & contractual cash flows | Connected | Real-time | 780K |
| Treasury Management System | HQLA, repo & money-market positions | Connected | 2 min ago | 210K |
| Market Data Feed | Yields, haircuts & collateral pricing | Connected | Real-time | 180K |
| RBI Regulatory Reporting Hub | LCR / NSFR submission & norms | Connected | 6 min ago | 4K |
| Counterparty & Depositor Registry | Concentration & relationship data | Connected | 10 min ago | 62K |
| Payment & Settlement Rails | Intraday liquidity movements | Syncing | In progress | 390K |
No rows match your search.
Workspace
Workspace nameKestrel Bank
EnvironmentProduction
Data residencyMumbai, IN · On-prem
API keysk-live-••••••••7f1c
Team access
Neha Kulkarni · Group TreasurerActive now
Arvind Seth · ALCO Secretary18m ago
Liquidity Bot · Service accountActive now