Skip to content
RFJ
🔍Transparency · Methodology

The Methodology

How this site forms its verdicts, grounds its AI synthesis, structures its audits, and selects its references. Published in full because methodology transparency is a trust multiplier.

Every page on this site involves judgment. The verdicts on /audit, the strategy ranking on /exit, the AI synthesis on /ask, the scholarly selection on /consensus — all involve choices about what to include, what to emphasise, and how to characterise. This page documents those choices.
Last reviewed2 June 2026Next review due2 September 2026Corrections log

The Six Pillars of Real Compliance

This is the framework used throughout /audit to evaluate Islamic finance products. It is derived from AAOIFI Shariah Standards 8 + 9 + 12 (Murābaḥah, Ijārah, Mushārakah) combined with the AMJA + IIFA deliberations on contemporary product structures. Not original to this site; codified here as a single framework for consistent application.

  1. Genuine ownership transfer. Does the provider actually take ownership of the asset (with associated risks), or is the ownership a paper formality that immediately transfers to the buyer?
  2. Real economic risk-sharing. In a Mushārakah, do both partners genuinely bear loss exposure? Or is one partner contractually shielded?
  3. Time-of-use rent, not time-of-money rent. In an Ijārah, is rent paid for use of a real asset (the building, the car) — or is it a thinly-veiled charge for the time value of the borrowed money?
  4. No tying to benchmark interest rates. Mark-ups and rentals should be priced on real-world variables — not simply pegged to BBSW, LIBOR, or RBA cash rate. (Practical note: many providers reference benchmark rates for convenience; the structural question is whether the underlying economic logic is independent.)
  5. No synthetic risk-elimination clauses. Clauses that guarantee the provider's return regardless of asset performance, or guarantee the customer's principal regardless of Mushārakah outcome, structurally replicate riba.
  6. Substance matches form. The most important pillar. The legal-form test must be paired with an economic-substance test. A Murābaḥah where the bank never genuinely owns the goods, or an Ijārah where the bank's exposure to asset risk is engineered to zero, passes form but fails substance.

How audit verdicts are formed

Each /audit entry walks the provider's documented contract structure through the six pillars above. The verdict is assigned by counting how many pillars the structure satisfies + how confident we are in our reading of the contract:

What we don't do: We don't grade based on a provider's marketing claims or Shariah board membership alone. A Shariah board is a positive signal but not a verdict. We attempt to read the actual contract or product disclosure where available, and we mark verdicts as lower-confidence when we can't.

The /ask AI synthesis prompt — published

Users on /ask interacting with the AI synthesis mode see summaries built by Claude from the corpus. The prompt that shapes those summaries is published below so users can judge whether the framing is what they want.

SYSTEM PROMPT (excerpt — full version in /app/lib/ai-ask.ts): You are a research librarian for "Riba-Free Journey" — a notebook on Islamic finance for Western Muslims. Your role is STRICTLY synthesis of cited scholarly positions. You are NOT a mufti and DO NOT issue rulings. For every answer: 1. Ground every claim in the retrieved corpus entries. Cite the scholar by name. Quote where possible. 2. Where scholars disagree, present BOTH positions fairly. Do not privilege one over another based on Riba-Free Journey's own judgments — the user can decide. 3. If the corpus is silent on a question, say so plainly. Do not invent positions. Do not extrapolate beyond what the cited scholars explicitly addressed. 4. Refuse to issue rulings of any kind. Always remind the user to verify with a scholar they trust for binding decisions. 5. When asked about a specific provider, refer to that market's /audit verdict + the methodology page, not your own reasoning. You are bound by the corpus. The corpus is bound by sourcing. Sourcing is bound by humility.

The full prompt + retrieval logic is in app/lib/ai-ask.ts in the public GitHub repository. The repository is open-source-spirit (we publish the source so the methodology is auditable; we ask you not to clone-and-deploy without scholarly review of the corpus).

Calculator assumptions — published

The five calculators on this site each make specific assumptions to produce point-estimate outputs. The assumptions are listed in each calculator's "Methodology" expandable section, and also catalogued here.

Where currency figures appear below, they are the Australian-edition default anchors. The calculators themselves are market-aware: choosing a US, UK or Canada edition re-anchors the currency, medians and thresholds to that market. The rates and ratios (returns, appreciation, transaction-cost percentages) are assumptions that hold across editions.

Exit calculator (/exit)

Halal housing calculator (/tools/halal-housing)

Zakat calculator (/tools/zakat)

Mawārīth calculator (/tools/mawarith)

Hijrah readiness (/tools/hijrah-readiness)

Citation standards

Across all pages, citations follow these conventions:

How stories are compiled

The 9+ stories on /stories are composite patterns, not individual case records. They are compiled from:

Each story includes a "Composite · pattern" tag and a sources line. Specific dollar figures and timelines are representative of the underlying pattern, not records of a single named family.

Update + correction policy

The site is a living document. When facts change (tax rates, property medians, provider products, scholarly positions), the content should change with them. When errors are found, they should be corrected publicly + transparently.

Last reviewed: 2026-05-27 · Methodology updates are published as the site matures.

Ask