Treasury & Liquidity Operations
The 4 screens that forecast cash runway, monitor FX exposure and hedging, manage counterparty credit limits, and track Basel III liquidity buffers.
What this group does
Treasury & Liquidity Operations covers cash-runway forecasting, currency exposure/hedging, counterparty credit limits, and Basel III liquidity buffers. All 4 screens use the identical direct-per-record two-stage hybrid engine (Stage 1 embedding infrastructure, unwired to the UI; Stage 2 AI-engine-routed synthesis) described on the Overview page.
Screen preview
A quick look at the screen. Use the toggle in the corner to switch between the dark and light themes.
Liquidity buffer, hedge coverage and debt maturity
1. Cash Flow Forecast
Purpose: A treasury ledger tracking an operating entity’s 30/60/90-day net cash projection against a targeted minimum reserve buffer — optimal, tight, or in deficit — so treasury teams can decide whether a liquidity sweep or credit-line draw is needed.
Ledger columns / key fields
- ID format
CFF-YYYY-XXXX forecastId,operatingEntity,netCash30d,netCash60d,netCash90d,projectedMinBuffer,currencyvarianceIndexPct,riskLevel(optimal/tight/deficit)countryCode,lastRunDate,aiStressTestAnalysis
Gauge
Net Cash Runway Gauge — the percentage of records at optimal risk level out of the total. Standard status-count pattern. Bar color: ≥80% green, 40–79% amber, below red.
Key actions
- New Record, AI Analyze (runs the cash flow forecast audit, producing a stress-test analysis plus confidence score; disabled while operating entity is blank).
- Execute Liquidity Sweep / Sweep (row action, inline mini-form: amount + source account + target account + reason) — logs the sweep intent to the notes and updates the last-run date; disabled when the risk level is deficit.
- Delete — permanent, with inline confirmation.
Known limitation
Sweep does not move real money between bank accounts. The app’s real Treasury module only covers term-deposit/bond/money-market-fund maturity and rollover — it offers no generic “transfer amount between two accounts” capability, and there is no link between this screen’s cash flow forecast records and a real treasury instrument or financial account. The action only logs intent to its own record.
2. FX Exposure & Risk
Purpose: A currency risk ledger tracking a currency pair’s gross asset/liability position, current hedge coverage, and daily Value-at-Risk (99% confidence) — fully hedged, partially exposed, or critically exposed.
Ledger columns / key fields
- ID format
FXR-YYYY-XXXX exposureId,currencyPair,grossAssets,grossLiabilities,baseCurrency,dailyVar99hedgeCoveragePct(0–100),auditStatus(hedged/partially_exposed/critical_exposure)countryCode,lastRunDate,aiHedgingRecommendation
Gauge
FX Hedge Coverage Bar — This is an average, not a status count: it takes the mean of every record’s own hedge-coverage percentage, since that field is already a per-record percentage. Bar color: ≥80% green, 40–79% amber, below red.
Key actions
- New Record, AI Analyze (runs the FX exposure audit, producing a hedging recommendation plus confidence score; disabled while currency pair is blank).
- Auto-Generate FX Hedge Order (row action, inline mini-form: notional amount + instrument type (forward/option) + target rate + reason) — logs the order intent to the notes and updates the last-run date; disabled when the audit status is hedged.
- Delete — permanent.
Known limitation
Hedge Order does not send a real order to a bank or broker. There is no FX trading/execution system anywhere in the app today; the Treasury module only covers term deposits/bonds/money-market funds, with no FX forward/option capability. The action only logs order intent to its own record.
3. Credit Risk & Limits
Purpose: A counterparty credit risk ledger tracking a counterparty’s approved credit ceiling against current drawdown — within limit, at the warning threshold, or breached.
Ledger columns / key fields
- ID format
CRL-YYYY-XXXX limitId,counterpartyName,entityRating,approvedCeiling,currentOutstanding,currencycomplianceStatus(compliant/warning_threshold/limit_breached)- Country code, last-run date, AI credit diagnostic. Headroom and utilization are derived: Headroom = approved ceiling minus current outstanding; Utilization = current outstanding divided by approved ceiling, times 100.
Gauge
Counterparty Concentration Meter — Not a status-count ratio: the largest single counterparty’s current outstanding balance divided by the sum of every record’s current outstanding balance, times 100. This measures concentration risk — how dominant the single biggest exposure is in the portfolio — not the share of compliant vs. breached records. Bar color: ≥60% red, ≥30% amber, below green.
Key actions
- New Record, AI Analyze (runs the credit risk audit, producing a credit diagnostic plus confidence score; disabled while counterparty name is blank).
- Recalibrate Credit Ceiling (row action, inline mini-form: new ceiling amount) — updates the approved ceiling.
- Delete — permanent.
Exception: this is a real, direct update
Unlike Sweep and Hedge Order above, Recalibrate Credit Ceiling has a genuine, direct effect on the record’s own data — the ceiling value lives entirely within this record, not in an external system. It directly updates the approved ceiling to the new value and appends the change (old → new) to the notes for audit trail. Headroom and Utilization automatically recompute against the new ceiling. This button is never disabled.
4. Liquidity Buffer
Purpose: The final screen in this group. A Basel III liquidity reserve ledger tracking an asset class or custodial institution’s HQLA (High-Quality Liquid Asset) reserve position against its statutory requirement. Each record carries a contribution percentage toward the portfolio-wide Liquidity Coverage Ratio (LCR).
Ledger columns / key fields
- ID format
LQB-YYYY-XXXX bufferId,assetClassOrInstitution,reserveTier(tier1_cash/tier2_sovereign_debt/tier3_money_market)marketValue,statutoryRequirement,currencystatus(sufficient/near_threshold/deficit),countryCode,lastRunDate,aiStressAnalysis
Gauge
Liquidity Coverage Ratio Gauge — Capital-weighted, not status-count: sum of every record’s market value divided by the sum of every record’s statutory requirement, times 100, shown to one decimal place, never rounded to an integer.
Critical nuance: values above 100% are valid and desirable
Unlike every other gauge in this add-on (where 100% means “all records in good standing”), the LCR gauge follows Basel III semantics: 100% is the regulatory minimum, not a ceiling. Values above 100% represent an over-reserved, strong position — not an error or a display bug. Color thresholds: <100% red (below the regulatory minimum, non-compliant); 100–119.9% amber (above minimum but thin margin, adequate); ≥120% green (strong reserve position). The bar’s visual width is capped at 100% so it never overflows its container, but the numeric percentage text next to it is never clamped — it always shows the real computed value (e.g. 142.0%). The same logic applies to each row’s own LCR Contribution column.
Key actions
- New Record, AI Analyze (runs the liquidity buffer audit, producing a stress analysis plus confidence score; disabled while asset class/institution is blank).
- Reallocate HQLA Reserves (row action, inline mini-form: target reserve tier) — moves the record to the new reserve tier.
- Delete — permanent.
Exception: this is a real, direct update
Like Recalibrate Credit Ceiling, Reallocate HQLA Reserves has a genuine, direct effect on the record’s own data — the reserve-tier assignment lives entirely within this record, not in an external system. It directly updates the reserve tier to the new tier and appends the change (old tier → new tier) to the notes. This button is never disabled.
AI Assistant — Cash Flow Forecast
The Cash Flow Forecast screen ships with an embedded AI Assistant — the same right-edge panel used across TotalApp. Open it from the vertical AI Assistant tab on the right of the screen; it understands what is currently on your screen and answers in Markdown, matching the language you write in (Turkish or English).
This chat assistant is a separate, complementary feature from the screen's existing per-record AI Analyze button — AI Analyze runs a fixed stress-test on one entity and returns an analysis for that record, while this assistant is an open-ended conversation across the whole portfolio that you can steer, ask follow-up and what-if questions in, and use to explore results AI Analyze already surfaced. It is tuned for portfolio-wide cash runway review:
- Forecast-risk narration — explains which entities are driving forecast risk across the portfolio.
- What-if scenarios — answers what-if questions about how a buffer change would play out.
- Below-minimum-buffer entities — flags entities projected to fall below their minimum reserve buffer.
- 30/60/90-day trend summary — summarizes how risk is trending across the 30/60/90-day windows.
Tick one or more rows in the table to scope the conversation, then use Analyze selected or Summarize selected. Any answer can be saved via Save as Report (it lands in My Reports), and Add Knowledge lets you attach reference documents to ground the reply.
Engine-aware
The assistant respects your Settings → Agentic writer engine: Ollama / local-LLM / in-browser models run entirely on your machine, while API routes through the server — the panel never leaks your data to a server the selected engine can’t reach.
AI Assistant — FX Exposure & Risk
The FX Exposure & Risk screen ships with an embedded AI Assistant — the same right-edge panel used across TotalApp. Open it from the vertical AI Assistant tab on the right of the screen; it understands what is currently on your screen and answers in Markdown, matching the language you write in (Turkish or English).
This chat assistant is a separate, complementary feature from the screen's existing per-record AI Analyze button — AI Analyze runs a fixed hedging analysis on one currency pair and returns a recommendation for that record, while this assistant is an open-ended conversation across every currency pair in the portfolio that you can steer, ask follow-up questions in, and use to explore results AI Analyze already surfaced. It is tuned for portfolio-wide hedge review:
- Hedge-coverage gaps — analyzes portfolio-level hedge-coverage gaps across currency pairs.
- High VaR, low coverage — flags pairs that combine high Value-at-Risk with low hedge coverage.
- Net exposure by currency — summarizes net exposure broken down by currency.
- Concentration risk — discusses concentration risk across currency pairs.
Tick one or more rows in the table to scope the conversation, then use Analyze selected or Summarize selected. Any answer can be saved via Save as Report (it lands in My Reports), and Add Knowledge lets you attach reference documents to ground the reply.
Engine-aware
The assistant respects your Settings → Agentic writer engine: Ollama / local-LLM / in-browser models run entirely on your machine, while API routes through the server — the panel never leaks your data to a server the selected engine can’t reach.
AI Assistant — Credit Risk & Limits
The Credit Risk & Limits screen ships with an embedded AI Assistant — the same right-edge panel used across TotalApp. Open it from the vertical AI Assistant tab on the right of the screen; it understands what is currently on your screen and answers in Markdown, matching the language you write in (Turkish or English).
This chat assistant is a separate, complementary feature from the screen's existing per-record AI Analyze button — AI Analyze runs a fixed credit diagnostic on one counterparty and returns a result for that record, while this assistant is an open-ended conversation across the whole counterparty portfolio that you can steer, ask follow-up questions in, and use to explore results AI Analyze already surfaced. It is tuned for portfolio-wide concentration review:
- Concentration risk — explains concentration risk across counterparties.
- Near-ceiling counterparties — flags counterparties approaching their approved ceiling.
- Compliance-status distribution — summarizes the compliance-status distribution by rating.
- Review prioritization — prioritizes which counterparties most need a limit review.
Tick one or more rows in the table to scope the conversation, then use Analyze selected or Summarize selected. Any answer can be saved via Save as Report (it lands in My Reports), and Add Knowledge lets you attach reference documents to ground the reply.
Engine-aware
The assistant respects your Settings → Agentic writer engine: Ollama / local-LLM / in-browser models run entirely on your machine, while API routes through the server — the panel never leaks your data to a server the selected engine can’t reach.
AI Assistant — Liquidity Buffer
The Liquidity Buffer screen ships with an embedded AI Assistant — the same right-edge panel used across TotalApp. Open it from the vertical AI Assistant tab on the right of the screen; it understands what is currently on your screen and answers in Markdown, matching the language you write in (Turkish or English).
This chat assistant is a separate, complementary feature from the screen's existing per-record AI Analyze button — AI Analyze runs a fixed stress analysis on one asset class or institution and returns a result for that record, while this assistant is an open-ended conversation across the whole reserve portfolio that you can steer, ask follow-up questions in, and use to explore results AI Analyze already surfaced. It is tuned for portfolio-wide liquidity review:
- Shortfall narration — narrates Liquidity Coverage Ratio and statutory-requirement shortfalls across asset classes or institutions.
- Reallocation suggestions — suggests reallocating reserves to close a shortfall.
- Adequacy by reserve tier — summarizes buffer adequacy broken down by reserve tier.
- Below-statutory flags — flags institutions sitting below their statutory requirement.
Tick one or more rows in the table to scope the conversation, then use Analyze selected or Summarize selected. Any answer can be saved via Save as Report (it lands in My Reports), and Add Knowledge lets you attach reference documents to ground the reply.
Engine-aware
The assistant respects your Settings → Agentic writer engine: Ollama / local-LLM / in-browser models run entirely on your machine, while API routes through the server — the panel never leaks your data to a server the selected engine can’t reach.
Architecture note: embedding layer present but not wired to the UI
All 4 screens follow the add-on-wide two-stage hybrid engine. Stage 1 (embedding) semantic-search infrastructure exists for each of the 4 screens (treasury, FX hedging, counterparty, and regulatory search respectively) — but none of the four screens use it yet. Users go straight from filling the form to pressing “AI Analyze,” which triggers Stage 2 (AI-engine-routed synthesis) directly. See the Overview page for the full architecture.