Forecast receivables collection and liquidity
Forecast cash collection, disputes, defaults and minimum liquidity from right-censored receivable histories: fit empirical-Bayes categorical transition probabilities by lawful aggregate risk class, state and age; retain censored exposure; simulate every current aggregate receivable under shared market/cash scenarios; and abstain on unsupported states, unverified evidence or inadequate liquidity probability.
What it's for
Turns receivables from a static aging table into a right-censored cash, dispute, default and liquidity distribution with explicit support and evidence refusal.
What you give it
Every tool is a pure function: it computes from what you pass it and reads nothing on its own. Inputs split into evidence you supply from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.
| Field | Type | Role | Required |
|---|---|---|---|
| current_receivables | array of objects (6 fields) | Evidence | Yes |
| current_unrestricted_cash | number ≥ 0 | Your calibration | Yes |
| historical_state_episodes | array of objects (7 fields) ≥ 10 items | Evidence | Yes |
| horizon_periods | integer ≥ 1, ≤ 60 | Your calibration | Yes |
| market_scenarios | array of objects (8 fields) | Evidence | Yes |
| max_detail_rows | integer ≥ 1, ≤ 500 | Numerical control | Optional |
| maximum_age_bucket | integer ≥ 1, ≤ 120 | Your calibration | Optional |
| minimum_liquidity_probability | number ≥ 0, ≤ 1 | Your calibration | Optional |
| minimum_state_exposure | number ≥ 0 | Your calibration | Optional |
| minimum_unrestricted_cash | number | Your calibration | Yes |
| prior_strength | number > 0 | Your calibration | Optional |
| seed | integer ≥ 0, ≤ 2147483647 | Numerical control | Optional |
| simulations | integer ≥ 100, ≤ 100000 | Numerical control | Optional |
Each market_scenarios
record
| Field | Type | Required |
|---|---|---|
| default_hazard_multiplier | number (≥ 0, ≤ 10) | Yes |
| default_recovery_fraction | number (≥ 0, ≤ 1) | Yes |
| dispute_hazard_multiplier | number (≥ 0, ≤ 10) | Yes |
| evidence_verified | boolean | Yes |
| fixed_cash_flow_by_period | array of number (≥ 1 item) | Yes |
| id | string (non-empty) | Yes |
| payment_hazard_multiplier | number (≥ 0, ≤ 10) | Yes |
| probability | number (≥ 0, ≤ 1) | Yes |
{
"current_receivables": [
{
"age_periods": 1,
"evidence_verified": true,
"face_amount": 100,
"id": "open-portfolio",
"risk_class": "standard-commercial",
"state": "open"
},
{
"age_periods": 2,
"evidence_verified": true,
"face_amount": 50,
"id": "disputed-portfolio",
"risk_class": "standard-commercial",
"state": "disputed"
}
],
"current_unrestricted_cash": 100,
"historical_state_episodes": [
{
"duration_periods": 2,
"evidence_verified": true,
"id": "receivable-episode-0",
"outcome": "paid",
"risk_class": "standard-commercial",
"start_age_periods": 0,
"state": "open"
},
{
"duration_periods": 2,
"evidence_verified": true,
"id": "receivable-episode-1",
"outcome": "disputed",
"risk_class": "standard-commercial",
"start_age_periods": 1,
"state": "open"
},
{
"duration_periods": 2,
"evidence_verified": true,
"id": "receivable-episode-2",
"outcome": "defaulted",Truncated for display — the full payload is 598 lines.
What you get back
This is the actual output of running the example above — computed by the same function the platform calls, not an illustration.
{
"configuration": {
"horizon_periods": 3,
"maximum_age_bucket": 12,
"minimum_state_exposure": 0,
"prior_strength": 5,
"seed": 3,
"simulations": 500
},
"decision": "receivables_collection_forecast_supported",
"failed_gates": [],
"guardrails": [
"Risk classes must be aggregate, lawful operating cohorts and never inferred protected or personal traits.",
"Right-censored episodes remain exposure; unresolved invoices are not defaults or zero cash.",
"Forecasted collections are conditional planning distributions, not audited cash, credit scores or permission for customer coercion."
],
"method": "empirical_bayes_age_state_competing_risk_receivables_simulation",
"state_support_diagnostics": [
{
"period_exposure": 60,
"risk_class": "standard-commercial",
"state": "disputed"
},
{
"period_exposure": 60,
"risk_class": "standard-commercial",
"state": "open"
}
],
"summary": {
"collected_cash_p10": 2.5,
"collected_cash_p50": 15,
"collected_cash_p90": 105,
"current_face_amount": 150,
"current_receivable_count": 2,
"defaulted_face_amount_p90": 150,
"disputed_face_amount_p90": 100,
"liquidity_probability": 1,
"minimum_cash_p10": 90
},
"truncation": {
"support_rows_omitted": 0
},
"unsupported_risk_states": []Truncated for display — the full payload is 45 lines.
How it works
Forecasting & survival — Estimate when something completes or fails, with censoring and unresolved work handled honestly rather than dropped.
- 1 Convert resolved and censored open/disputed state episodes into period exposures and paid, disputed, defaulted or survived outcomes without relabeling censoring as default.
- 2 Estimate age-state transition probabilities with risk-class pooled empirical-Bayes shrinkage, including prior-only age cells inside the governed range, and gate current states on total historical exposure.
- 3 Draw common market scenarios, simulate each current receivable through open/disputed/paid/defaulted states, credit paid and default-recovery cash at event time, add coherent fixed cash flows, and report collection/default/dispute and minimum-cash distributions.
Before you trust it
Every tool in the catalog ships with the conditions under which its answer is meaningful — and the conditions under which it should abstain instead of guessing.
Assumptions & guardrails
- Training examples precede their outcomes, censoring and unresolved work are represented, and deployment populations remain comparable to validation cohorts.
- State/outcome taxonomy, cadence, age origin, censoring, risk classes and collection process are stable; current cohorts have historical support; scenario hazard multipliers and recovery fractions are coherent; invoices and cash flows share a currency and perimeter.
- A predictive interval or risk estimate is not a deadline promise, causal explanation, or individual-performance judgment.
- This is aggregate liquidity planning, not a customer credit score, audited cash forecast or permission for coercive collection. Protected traits, named-person behavior and LLM-inferred risk classes are prohibited.
Minimum evidence
- historical_state_episodes: at least 10 rows/items
- current_receivables: required and organization-defined
- market_scenarios: required and organization-defined
- current_unrestricted_cash: required and organization-defined
- minimum_unrestricted_cash: required and organization-defined
- horizon_periods: required and organization-defined
How to validate it
Use chronological train/calibration/test windows, compare proper scores and decision value with a simple baseline, and recalibrate only from outcomes resolved after prediction time.
Calibrating it to your org
Same for everyone
The mathematical kernel, validation rules, method version, and JSON output semantics are organization-independent; no tenant-trained coefficients or company benchmark is embedded in the function.
Specific to you
- version-matched invoice event histories transformed into non-overlapping state episodes retaining censored exposure, joined to a point-in-time aggregate receivable snapshot and finance-owned common cash scenarios
- invoice and state taxonomy, risk-class lawfulness/privacy, age origin and cadence, censoring/maturity, payment/dispute/default/recovery definitions, process epoch, scenario dependence, cash perimeter, age/prior/support controls, simulations and liquidity gate
Calibration workflow
- 1 Define the management decision, target outcome, aggregate unit, privacy boundary, cadence, and prediction/intervention horizon for this organization.
- 2 Build a tenant-scoped historical cohort using only information available before each prediction or decision; preserve zero periods, censoring, assignment probabilities, and unresolved outcomes when the method requires them.
- 3 Estimate statistical parameters on training history, but obtain costs, utilities, risk tolerance, practical-effect thresholds, capacity, and policy constraints from accountable decision owners.
- 4 Validate on later time windows or held-out aggregate units at the deployment grain, against a simple baseline and the function-specific validation strategy.
- 5 Deploy only if the returned decision clears evidence, overlap, calibration, robustness, and guardrail checks; warning, unsupported, schema-gap, and fallback decisions are abstentions.
- 6 Monitor realized outcomes, data drift, coverage, and decision regret; recalibrate at a governed cadence or after a detected regime/definition change, never merely because a stakeholder dislikes the result.
Call it from your AI
You don't wire up 388 tools in your MCP client. The GitRevio MCP server exposes 20 tools, three of which let an agent search the catalog, read a tool's schema, and run it — so the assistant finds this one on its own.
gitrevio_capabilities_search
{ "q": "forecast cash collection disputes defaults and" }
→ finds "forecast_receivables_collection_and_liquidity"
gitrevio_capability_describe
{ "capability_id": "forecast_receivables_collection_and_liquidity" }
→ returns the input schema and agent guidance shown on this page
gitrevio_capability_run
{ "capability_id": "forecast_receivables_collection_and_liquidity", "arguments": { ... } }
→ returns the result shown aboveWorks in Claude Desktop, Claude Code, Cursor, Cline, Continue.dev, Goose and Aider. See the MCP server.