Pelican Cortex

The model proposes.The rules decide.

Sovereign, auditable, quantum-ready AI for financial crime prevention. Built to move banks from legacy systems to a defensible future.

0+
years of proprietary AI expertise
0+
countries with Pelican customers
0B+
transactions processed
$0T+
in transaction value

Why Cortex

From legacy to a defensible future

Pelican’s customers run on proven systems today. Cortex is how we move them into a future they can defend to regulators, to their boards and to the next decade of threats: faster, more capable and more secure.

01Speed  ·  Seconds

Legacy was built for batch. Payments now move in seconds.

Legacy

Rules-era screening is accountable but noisy. Very high false-positive rates leave analysts spending most of their time clearing legitimate payments.

With Cortex

AI triage clears false positives, and decisions are designed to land inside the payment window on real-time rails.

02Capability  ·  May 2027

Modern AI has to be explainable before banks can use it.

Legacy

Black-box AI is accurate but unaccountable. Banks cannot explain or audit its decisions to regulators or their own model risk teams, so it stays out of production.

With Cortex

Neural models propose and a symbolic engine, compiled from regulation, decides. Every decision traces to a rule, and every rule to a clause of law. OSFI Guideline E-23 on model risk takes effect in May 2027.

03Security  ·  2030–35

Today’s security has to outlast today’s cryptography.

Legacy

Evidence signed with today’s cryptography stops being proof if a signature can be forged later. Data sent to outside AI services also leaves the bank’s walls.

With Cortex

Evidence records designed to be sealed with post-quantum signatures, running on-premise or air-gapped with no outside AI model APIs. The G7 roadmap points critical systems to quantum-safe cryptography by 2030–32 and the sector by 2035.

Sources: OSFI Guideline E-23; G7 Cyber Expert Group post-quantum roadmap, January 2026.

Proven today

A great addition to a proven suite

Pelican has spent 30+ years building AI for payments and financial crime compliance, and our customers rely on it in 55+ countries. Cortex joins five proven solutions. Modular, not bundled: each runs on its own or combined, on-premise or in the cloud.

ESSLO

Enterprise Sanctions Self-Learning Optimization

Layers on top of your existing screening engine with no rip-and-replace, and learns from analyst behaviour over time.

70% average false positive reduction
iCaaS

Intelligent Compliance as a Service

A cloud platform for transaction and name screening, AML monitoring, KYC and fraud detection, all in one place.

80% alert review time reduction
FAH

Financial AI Hub

Enriches, repairs and intelligently routes payments to lift straight-through processing, and turns payment data into analytics.

98%+ STP reported by users
iPaaS

Intelligent Payments as a Service

An API-driven payments platform for high-value, low-value, instant and cross-border payments, with compliance built in.

New rails without a core overhaul
Pelican Trade

Trade-based money laundering solution

Automates trade document review, sanctions and dual-use goods screening, and TBML red-flag detection across the trade cycle.

Explainable TBML detection
New  ·  Pelican Cortex

The model proposes. The rules decide.

Sovereign, auditable, quantum-ready AI for financial crime prevention, with a replayable evidence record for every decision.

Quantum-ready by design

Figures as stated in Pelican product materials.

How Cortex decides

How Cortex makes a decision

1

Payment arrives

From domestic rails and international formats, in batch or in real time.

2

Neural layer proposes

A financial crime language model and behavioural and network models read the payment and score risk.

3

Symbolic engine decides

Regulation compiled into verified rules makes the final call on every payment.

4

Evidence is sealed

A tamper-evident record captures the decision so it can be replayed exactly.

When the two layers disagree, the stricter outcome wins.

The product

Three engines on one platform

Shared by all three: one knowledge graph, one policy engine, one evidence ledger.

01

Payment and sanctions screening

Semantic name matching across scripts and aliases. Deterministic sanctions logic, including ownership rules. AI triage of false positives.

02

Anomaly detection

Behavioural and network models that find fraud, money mule rings and laundering patterns, including on real-time rails.

03

Global compliance

Regulation turned into formal rules. Each rule is checked by a mathematical solver and signed off by a named compliance officer before it runs.

How we bring it to you

Start beside what you run. Expand when it is proven.

Cortex does not ask a bank to rip and replace. It is designed to begin next to your existing screening and to grow only as each stage earns its place.

STAGE 1

Start beside what you run

Cortex runs in sidecar mode alongside your existing screening, so live payments carry no added risk. It is the same deployment model Pelican’s screening optimiser already uses at a tier 1 global bank.

STAGE 2

Prove it on your own data

Results are measured against your own baseline, including false positives and analyst effort. Each stage starts only after the one before it has been proven.

STAGE 3

Extend across the platform

Anomaly detection for fraud and mule networks, and the regulation-to-rules pipeline, are added with model risk approval at your institution.

STAGE 4

Move inline for instant payments

Decisions are made inside the payment window, on the same rules and the same replayable evidence you have already approved.

This is Pelican’s planned approach. The sequence and timing will depend on each institution and on technical confirmation.

Auditability

Every decision can be replayed for an examiner

Cortex writes a tamper-evident evidence record for each decision. It can be replayed exactly, with the model, rule and list versions in force at the time. Each rule links back to the clause of law it enforces.

In every record

◆  The payment as it was received
◆  The model version and its score
◆  The rule version and its legal source
◆  The watchlist version
◆  The officer who approved the rule

Quantum-ready by design

An evidence record sealed today must still stand in 2035

Audit evidence outlives the cryptography that signed it. If a signature can be forged later, the record stops being proof.

01  SEAL

Signed twice

Each record is designed to be signed with classical and post-quantum signatures together.

02  RE-SEAL

Built for change

As standards change, records are re-sealed and the chain of custody is kept.

03  VERIFY

Provable later

An examiner can check any record, whenever it was written.

Theme
What the banks are signalling
How Cortex is designed to answer
Quantum-safe security
Client services move to post-quantum cryptography
Evidence records sealed with post-quantum signatures from the first release
Crypto-agility
Be ready to change algorithms again
Signing and key exchange are configuration, and each record names its scheme
Third-party readiness
Vendors sit inside the migration scope
Runs on-premise or air-gapped, under the bank’s own key management
Quantum-enabled solutions
Explore quantum for hard problems
Network and solver workloads sit behind interfaces, ready to test new compute

The largest institutions have started the quantum clock

RBC · September 2026

A bank puts a leader and a roadmap behind quantum

First Director of Quantum, two university partnerships, and a Quantum-Safe Security Program built on post-quantum cryptography and quantum key distribution.

G7 Cyber Expert Group · January 2026

The sector gets a shared timetable

A coordinated roadmap to post-quantum cryptography for banks, payment processors and their technology vendors, with 2035 as the sector-wide target.

G7 central banks · May 2026

Quantum becomes a governance practice

A first report on quantum technologies in finance, from a working group co-chaired by the Bank of Canada: start with visibility, build for agility.

Sources: RBC news release, 28 Sept 2026; G7 Cyber Expert Group roadmap, Jan 2026; G7 central banks report, May 2026. Quantum capabilities describe design intent and remain subject to technical confirmation.

Sovereignty

Runs where the bank’s data already lives

The models run inside the deployment, with no dependency on outside AI model APIs.

On-premise

Inside the bank’s own data centre.

Air-gapped

Fully disconnected, for the most sensitive environments.

Private cloud

In-region, starting with Canadian hosting.

Designed to connect toLynxACSS / AFTInteracReal-Time RailISO 20022SWIFT

Learn more about Pelican

Explore our solutions for payments, financial crime compliance and trade finance.

Visit our main website

© 2026 Pelican AI Corp. (CSE: PEL | FSE: E6V | WKN: A42LW1). All rights reserved.

Hinweis: Die nachfolgenden rechtlichen Hinweise gelten in ihrer englischen Fassung.

This page contains forward-looking information regarding the design, capabilities, intended deployment and timing of Pelican Cortex, including its intended use of post-quantum cryptography. Such information is subject to risks and uncertainties, including technical development and testing outcomes, and actual results may differ materially. See the Company’s public disclosure on SEDAR+. The CSE has not reviewed and does not accept responsibility for the adequacy or accuracy of this content.