Decision intelligence infrastructure.
Thalra evaluates proposed investment actions against financial context and versioned institutional policy, then returns explainable outcomes with evidence your systems can store and audit.
Evaluation architecture
How context becomes a governed outcome.
Platforms send customer, portfolio, behaviour and action context. Thalra derives signals, applies the active policy pack, and returns a decision with structured evidence — deterministic rules, not opaque scores.
- 01
Context
Intake financial observations for a proposed action — and surface what is missing or stale.
- 02
Signals
Derive material observations such as concentration, liquidity impact and recent risk escalation.
- 03
Policy
Apply versioned institutional policy packs with deterministic, named rules.
- 04
Decision
Return ALLOW, SUPPORT, WARN, CONFIRM or REVIEW — outcomes systems can act on.
- 05
Evidence
Attach rule identity, pack version, engine version and supporting detail.
Capabilities
Context intake & readiness
ImplementedAssess which observations are available, derived, unavailable or stale — so missing data is not mistaken for a clean pass.
Financial & behavioural signals
ImplementedMaterial observations such as concentration, liquidity impact, risk-profile mismatch and recent escalation — always distinct from the final decision.
Institutional policy evaluation
ImplementedVersioned policy packs with named rules that bind to ALLOW, SUPPORT, WARN, CONFIRM or REVIEW.
Evidence
ImplementedEvery response includes rule identity, pack version, engine version and supporting signal detail for audit and explanation.
Temporal analysis
ImplementedCompare earlier and updated financial states for the same proposed action — when context changes, outcomes can change.
Decision Relevance research
Research onlyExploring whether missing inputs would change a decision under bound ranges. Not a validated production intelligence product.
Context readiness
Know what can be evaluated.
Sparse institutional data is common. Thalra distinguishes observation status so platforms understand eligibility before treating an outcome as complete.
Context map
available · derived · unavailable · stale
- available
Customer risk profile
Observed
- available
Portfolio holdings
Observed
- available
Proposed action
Observed
- derived
Derived exposure impact
Evaluable
- unavailable
Liquidity impact signal
Needs liquid assets
- stale
Behaviour window
Freshness insufficient
Example evaluation path
Decision trace
Illustrative synthetic example
Context
Observed
Risk profile · portfolio · behaviour · proposed action
Signals
4 material
Concentration · exposure · mismatch · recent escalation
Policy
Applied
Versioned institutional rules
Decision
CONFIRM
Governed outcome for the proposed BUY
Evidence
Returned
Rule · policy pack · engine identity