Data Flow
Silsilat’s data flow is designed for transparency, traceability, and real-time settlement. Each event in the ecosystem — from loan creation to final redemption — passes through a verifiable pipeline that links off-chain data (e.g., gold appraisal, AML check) to on-chain consensus and liquidity execution.
Below we break down the end-to-end process across five major stages, each reinforced by Hedera Consensus Service (HCS), IPFS artifact creation, and Arize Phoenix observability.
Pawnshop Onboarding & Collateral Tokenization
Process Narrative
When a pawnshop or Ar-Rahnu outlet joins the Silsilat network, it is registered as an on-chain identity in the Silsilat Registry (HCS-11). Each gold receipt ("Surat Akuan Gadaian") is tokenized as a SAG Token with immutable metadata.
Sequence
Sample Metadata Schema
Key Points
Tokenization generates a verifiable digital twin of physical gold.
Each SAG token is non-transferable until verified and evaluated.
HCS anchoring ensures authenticity before valuation.
AI-Driven Gold Evaluation
Process Narrative
The Gold Evaluator Agent calculates the real-time value and permissible Loan-to-Value (LTV) ratio using live data and regulatory policy packs.All inputs, outputs, and model metadata are stored as trace artifacts and auditable via Arize Phoenix.
Sequence
Example Artifact (stored in IPFS)
Key Points
Each evaluation has a trace ID linking Phoenix logs, IPFS artifact, and HCS event.
LTV ratios and haircuts are policy-bound and automatically verified.
Administrators can later audit or override via
HCS_OVERRIDE_TOPIC_ID.
Liquidity Pool Execution
Process Narrative
Once the LTV and valuation are approved, the pawnshop requests liquidity settlement. The Silsilat Liquidity Pool instantly buys the SAG token from the pawnshop, releasing funds. Investors deposit capital into the pool and receive Liquidity Tokens (LQT) as proof of participation.
Sequence
Example HCS Message
Key Points
Liquidity provisioning is instant and deterministic, governed by smart node logic.
Pool investors earn profit-share or yield based on policy-defined spreads.
All fund flows are traceable and reversible under regulatory oversight.
Compliance, Audit & Override
Process Narrative
Every transaction triggers the Policy Agent to validate KYC, AML, and Shariah criteria. If an issue arises (e.g., flagged borrower, unusual transaction size), an override is initiated.
Sequence
Example Override Record
Key Points
Overrides are digitally signed and publicly verifiable.
Regulators maintain full visibility of all decision pathways.
Each override links to an immutable Phoenix trace and IPFS record.
Maturity, Redemption & Settlement
Process Narrative
Upon loan maturity, the pawnshop redeems the pledged gold by repaying the principal and profit. The liquidity pool automatically buys back LQT tokens, burns the SAG token, and closes the trace.
Sequence
Example Final Settlement Record
Key Points
Settlement closes the trace lifecycle on HCS.
Investors automatically receive yield payouts.
Regulators and auditors can query all historical data using the trace ID.
Summary of Data Objects
Data Object
Storage
Purpose
SAG Token
Hedera Token Service
Represents gold-collateralized asset
Trace Artifact
IPFS
Captures AI model inputs/outputs and compliance proof
HCS Message
Hedera Consensus
Anchors immutable event records
Phoenix Trace
Arize Phoenix
Observability of model decisions
Policy Record
Policy Engine
Defines rules for valuation, AML, Shariah
LQT Token
Liquidity Pool
Represents investor participation
Trust Invariants Across the Flow
No event exists without a trace. Every evaluation or liquidity transaction must have a Phoenix trace ID.
Every trace has an artifact. Each inference is logged as an IPFS bundle with policy hash.
Every artifact is anchored. HCS ensures consensus-level immutability.
Every policy can be verified. Regulators can re-run compliance checks using the stored schema and trace inputs.
Last updated