Methodology

How the Shortfall Resolution Optimizer works

Graph provides facts. Engines calculate. AI explains.

Last reviewed
Reviewed by
Rostislav Sikora
Next review due

From situation to candidates

The decision begins with required amount, available amount, shortfall, purpose, deadline and jurisdiction. The Canadian Shortfall Resolution Graph supplies evidence-backed facts. A deterministic generator creates unranked candidates across non-debt, assistance, arrangement, deferral, existing-credit and new-borrowing categories where supported.

Plan A, Plan B and Plan C

The Optimizer can combine partial candidates. It seeks deadline-feasible coverage, then minimizes new debt, cost, repayment pressure and risk while maximizing confidence. Plan A is ranked first by the locked policy; Plan B is an alternative; Plan C is a fallback where justified. A missing feasible plan remains a valid result.

Optimizer methodology: optimizer-decision-model-v1.0.

Coverage certainty

Confirmed, conditional and pending coverage are separate. Conditional coverage is never described as guaranteed.

Financial certainty

Verified, estimated, range and unknown money states remain explicit. Unknown is never converted to zero.

Robustness

Robustness describes dependency sensitivity, not creditworthiness. It does not alter plan mathematics.

What-If

A changed situation is sent through the same Optimizer. It creates a recomputed decision and successor receipt; it does not edit the prior receipt.

Deterministic authority

Candidate generation, plan ranking, robustness, fallback, actions and receipts use deterministic code and approved structured facts. Generative AI is not used as current financial decision authority and cannot invent financial facts.

View the methodology changelog