You've seen the homepage. Here's the depth underneath it — the detection engine, the AI that acts, the supplier vault, the payment controls, the cryptographic evidence chain and the platform-level bodyguard that keeps it all safe.
Connect your inbox or accounting system once. RILLA checks every invoice, bank-detail change and supplier message before payday — and texts you the moment something's off. Manual scan is always there when you want to check something yourself.
Runs every 5 minutes across Gmail, Outlook, Xero, MYOB and forwarded SMS. Flags scams the instant they land — before you open them.
Tap Scan Now anytime — paste an email, drop a PDF, snap a photo or check a phone number. Full risk verdict in seconds.
A live agent watches every supplier, bank detail and recent scan, surfacing the risks that need action before you sit down to pay.
RILLA doesn't rely on a single signal. Twenty-one deterministic layers — from ABN/GST validation to domain reputation to behavioural drift — feed the AI reasoner, so a verdict is always backed by explainable evidence, not a black box.
SMS → email → invoice → ABN/BSB → ABA → pay-run hold. The only engine covering the entire payment fraud chain in one pass.
Each scan runs across multiple AI models. You see their level of agreement — unanimous, majority or split. Split verdicts force a mandatory verify step.
RILLA learns each supplier's habits — invoice naming, GST patterns, sending cadence. When behaviour drifts, you get a high-severity alert before the money moves.
Every scanned invoice builds a technical fingerprint — metadata, structure, signature patterns. After three documents, forged or re-templated invoices get caught automatically.
The four defences that stop the scams that slip through the gaps — account takeovers, forged invoices, and fake batch payments.
Store trusted supplier bank details in a locked vault. Any change triggers an SMS OTP to your verified mobile — a stolen password alone can't repoint a supplier to a mule account.
Before you pay a new or changed supplier, we text their known-good mobile with the BSB/account details. A 'NO' reply fires an instant critical alert — stopping social-engineering scams before money moves.
Upload your ABA payment batch and cross-check every destination account against the vault, known-good suppliers and risk signals before the pay-run leaves your bank.
RILLA remembers every supplier you've paid. If their bank details change between invoices, you get a high-severity alert — the #1 signal of business email compromise.
Beyond detection — RILLA's AI layer acts. It reconciles pay runs, interrogates risk, traces attack chains and clusters fraud rings, then writes a board-ready brief every week.
AI-driven Pay / Verify / Hold verdicts on every ABA batch, reconciling against the vault, supplier history and network signals — automatically, before the batch is released.
Ask any scan 'why?' in plain English. The interrogator walks you through every signal, every layer and the model agreement behind the verdict.
Reconstructs the full path an attack took — from the first spoofed email to the bank-detail swap — so you can see exactly how it happened and close the gap.
Clusters related scams across the RILLA network — shared mule accounts, lookalike domains, repeated invoice templates — to expose coordinated fraud rings before they reach you.
Every week RILLA writes a board-ready CFO narrative — dollars protected, top risks, pay-run verdicts and recommended actions — auto-generated and delivered.
All controls feed a single release verdict. Nothing fights itself. Critical risks auto-freeze the pay-run; a break-glass override exists for emergencies, drilled and logged.
One gate decides whether a payment releases. Every control — vault, fingerprint, network, AI — feeds it as a signal. No conflicting alerts, no decision paralysis.
One tap freezes every outbound payment across connected accounts. Releasing requires verified re-authorisation — your emergency stop for a suspected breach.
Before a batch leaves your bank, RILLA verifies every payee against the vault, the network and the AI verdict — and signs the run with a cryptographic attestation.
Banks and underwriters ask for proof a verdict existed, unaltered, on the date recorded. RILLA's tamper-evident hash chain and daily signed Merkle roots deliver exactly that — provable even against RILLA itself.
Every audit event is chained with SHA-256 and keyed HMAC. Any later edit or insert breaks the chain and is flagged. A stolen DB password still can't forge a valid row.
Every night at 00:20 UTC the day's audit records fold into a signed Ed25519 Merkle root. Anchors are public and verifiable — proof a verdict existed, unaltered, on its date.
Every scan exports as a signed PDF evidence pack formatted for CommBank, NAB, Westpac and ANZ fraud-team intake — verified facts, flags and the full AI reasoning trail.
A request-level gatekeeper that inspects every incoming request for injection, path traversal and honeypot probes, auto-bans hostile sources, and runs alongside MFA, key rotation and the network trust layer.
Inspects every request for SQL/XSS injection, path traversal and honeypot probes. Auto-bans hostile sources (600 req/min or signature match) at the edge.
Every account supports two-factor auth, with step-up verification on sensitive actions — bank-detail changes, pay-run releases and admin overrides.
Every scan shows how the wider network sees this payee — trusted, known, new or flagged — using hashed signals only, never PII. Dozens of safe payments = trust.
Any AI assistant that speaks Model Context Protocol — Claude, ChatGPT, Cursor — can call RILLA's fraud checks directly. RILLA inside the tools you already use.
Data stays on Australian infrastructure. Full audit trail, signed forensic reports on request, and a trust pack open for your bank, insurer or auditor to review.
Connect once and the whole stack runs automatically — or scan a single invoice right now, free.
© 2026 Rilla Shield Defence Pty Ltd · ABN 80 678 733 950 · Content protected by trademark — see Copyright & IP Notice.
