How Telstra and Quantium Are Rewriting Australia's Anti-Fraud Playbook

A deep-dive into the Quantium-Telstra joint venture — from Scam Indicator to Fraud Indicator — and how telco intelligence is reshaping fraud detection across Australian banking, government, and digital services.

An eighteenth-century painting of a crowd hauling a giant white wooden horse on wheels through the gates of Troy, watched from the city walls.
Feature image: Giovanni Domenico Tiepolo / Public domain / Wikimedia Commons.

Overview

Australia lost $3.08 billion to scams in 2024 — a 16% increase from the previous year, with Australians losing more money to scams than in any recorded period. Against this backdrop, a partnership between two of Australia's most consequential technology organisations — Telstra and Quantium — has produced what may be the most impactful anti-fraud innovation of the decade.

Through their joint venture Quantium Telstra (QT), the companies have developed two production-grade fraud detection systems: Scam Indicator, launched in 2023, and Fraud Indicator, launched in early 2025. Together, these systems represent a fundamental architectural shift — using telco intelligence as a fraud signal that financial institutions and digital services previously had no access to.

This article examines the partnership's architecture, real-world impact, expansion strategy, and why it matters for every organisation that processes customer identities or financial transactions in Australia.


The Problem: A Blind Spot in Australian Fraud Detection

The Scale of the Crisis

Australian scam losses have been accelerating for years. The ACCC's Scamwatch recorded $3.08 billion in reported losses during 2024, while the ABS estimated that nearly 200,000 Australians experienced identity theft in FY23 alone. The total economic cost — including labour, legal, and recovery expenses — extends far beyond direct financial losses.

The fundamental problem is asymmetry: scammers know more about their targets than the institutions trying to protect them. A fraudster can obtain a victim's identity documents, open accounts in their name, and complete phone verification using a device they control. Banks see the transaction but lack context about whether the person initiating it is legitimate or a sophisticated attacker operating from a device with suspicious usage patterns.

The Telco Intelligence Gap

Telcos sit on a rich, underutilised data asset: real-time intelligence about how mobile devices are being used. When a fraudster signs up for a new bank account using a freshly provisioned SIM card — one that receives no personal calls, sends no messages, and immediately starts receiving authentication codes from multiple services — that pattern is invisible to the bank but obvious to the telco.

Before 2023, this intelligence existed in isolation. Telstra knew about suspicious device behaviour; banks knew about suspicious account activity. Neither could see what the other saw.


Quantium Telstra: The Joint Venture

Structure and Purpose

Quantium Telstra (QT) is a joint venture between Quantium, Australia's leading data analytics company, and Telstra, the country's largest telecommunications provider. The venture was established specifically to productise telco intelligence for fraud prevention — transforming raw network data into actionable, privacy-preserving fraud signals.

Sandy Cameron, CEO of Quantium Telstra, has been the public face of the venture, articulating its mission: "The commercial impact of fraud and scams on Australian enterprises is alarming, and the full cost goes well beyond the money lost — including labour, legal and recovery costs."

The Data Architecture

At its core, the QT system operates on a simple but powerful principle: mobile service usage patterns differ dramatically between legitimate customers and fraudsters. The system analyses behavioural signals from Telstra's network — not content, not conversations, not location data — but metadata about how a mobile service is being used:

  • SIM provisioning patterns (freshly activated vs. long-standing)
  • Call and SMS volume anomalies
  • Device switching frequency
  • Authentication code reception patterns
  • Service interaction timing

These signals are processed through machine learning models trained on millions of known fraud and scam cases, producing a risk score that can be shared — in real time — with partner organisations via API.


Scam Indicator: The First Wave (2023)

How It Works

Scam Indicator was the first production deployment of QT technology, launched in collaboration with Commonwealth Bank (CommBank) in 2023. It addresses the most common scam vector in Australia: phone-based social engineering, where a fraudster tricks a customer into transferring money during an active phone call.

The system works through a real-time API integration:

  1. A CommBank customer initiates a transaction that triggers the bank's internal scam detection systems
  2. As part of its detection process, CBA calls the Telstra API to check whether the customer is currently on an active phone call
  3. If the customer is on a call — the primary indicator that a scam may be occurring — CBA can intervene: contacting the customer through alternative channels, applying additional verification, or temporarily holding the transaction
  4. The API returns only specific data points relevant to scam prevention. CBA does not have access to any underlying customer data, call content, or metadata beyond the scam-relevant signal

The Scam Indicator API interaction proceeds as follows:

sequenceDiagram
    participant C as Customer
    participant B as CommBank
    participant T as Telstra API
    C->>B: Initiates transaction
    B->>T: Check if customer is on active call
    T-->>B: Call status (yes/no)
    alt On an active call
        B->>C: Intervene via alternative channels
    end

Impact and Results

The proof-of-concept simulations projected that Scam Indicator could double the success rate of phone scam detection and prevention for joint Telstra-CommBank customers, potentially mitigating losses of $15–20 million per annum. Since going live, the technology has:

  • Helped protect CBA customers from losing millions of dollars to scammers
  • Been expanded to cover landlines, protecting the most vulnerable elderly customers
  • Contributed to a 70% reduction in customer scam losses at CommBank (as reported by CEO Matt Comyn)
  • Supported a nationwide decline in scam losses of 13.1% from 2022 to 2023 (per ACCC data)

Why It Matters Architecturally

Scam Indicator demonstrated that cross-industry collaboration — telco + bank — could produce fraud detection capabilities neither organisation could build alone. The system didn't require either party to share sensitive customer data; it simply created a bridge between two isolated intelligence domains.

This architectural pattern — privacy-preserving signal sharing rather than raw data sharing — is what makes the QT model both effective and regulatory-compliant.


Fraud Indicator: The Second Wave (2025)

The Evolution

If Scam Indicator addressed ongoing scams, Fraud Indicator targets pre-transaction identity theft — the process by which fraudsters establish fraudulent accounts using stolen identities before any money moves.

Launched in February 2025, Fraud Indicator represents a significant architectural evolution:

Aspect Scam Indicator Fraud Indicator
Threat vector Phone-based social engineering during transactions Identity theft and fraudulent account opening
Timing Real-time intervention during active scam Pre-emptive risk assessment at onboarding
Data signal Is the customer currently on a phone call? Does this mobile number exhibit fraud-like usage patterns?
Integration point Transaction approval workflow Customer onboarding / KYC process
Impact metric Doubled phone scam detection rate 25%+ increase in identity theft detection

How Fraud Indicator Works

The system operates at the point of customer onboarding:

  1. A prospective customer provides a mobile number during account opening (for passcodes, authentication, and verification)
  2. CommBank's Fraud Indicator API queries Telstra's network for intelligence about that mobile service's usage patterns
  3. The system evaluates whether the mobile service exhibits behaviours consistent with fraud — freshly provisioned SIMs, unusual authentication code volumes, atypical service interaction patterns
  4. If the risk score exceeds a threshold, the bank can trigger additional verification steps or flag the application for manual review

The Fraud Indicator API interaction proceeds as follows:

sequenceDiagram
    participant C as Prospective Customer
    participant B as CommBank
    participant T as Telstra API
    C->>B: Provides mobile number during onboarding
    B->>T: Query usage patterns for mobile number
    T-->>B: Risk score
    alt Score exceeds threshold
        B->>C: Additional verification steps
    end

The 25% Impact

Quantium Telstra projects that Fraud Indicator will increase CommBank's detection rate for fraudulently opened accounts by over 25% for joint customers. This is a significant margin — it means one in four identity theft attempts that would have previously succeeded is now intercepted.

Identity Theft: The Bigger Picture

The ABS estimated that nearly 200,000 Australians experienced identity theft in FY23 alone. Unlike phone scams, which tend to be one-off events, identity theft has long-term consequences: damaged credit scores, stolen funds, and emotional distress that can persist for years. Fraud Indicator addresses this at the point of entry — before the fraudulent account is even created.


The Architecture: Privacy-Preserving Signal Sharing

Design Principles

The QT system is built on several foundational principles that make it both effective and legally defensible:

1. No Raw Data Sharing

Partners never receive Telstra's underlying customer data. The API returns only specific, fraud-relevant signals — a risk assessment derived from telco intelligence, not the raw data itself. This eliminates privacy concerns and simplifies regulatory compliance.

2. Purpose-Bound Access

CBA (and future partners) can only access data points directly relevant to scam or fraud prevention. There is no surveillance capability, no general-purpose customer profiling, no secondary use of the data.

3. Real-Time API Integration

The system operates in milliseconds — fast enough to integrate into live transaction approval workflows without adding perceptible latency. This is critical: fraud detection only works if it happens before the damage occurs.

4. Machine Learning at the Edge

The ML models that produce risk scores run on Telstra's infrastructure. Partners receive the output (a risk signal), not the model or its training data. This protects intellectual property and prevents adversarial manipulation.

Technical Flow

The following diagram shows how a fraud query flows through the system.

flowchart LR
    P["Partner (CBA)"] -->|Initiates Query| A["QT Signal API"]
    A -->|Validates Request| T["Telstra Network"]
    T -->|Mobile Usage Patterns| M["ML Models"]
    M -->|Risk Score| A
    A -->|Risk Signal| P
    P --> BL["Partner applies business logic"]

Expansion Strategy: Beyond CommBank

Multi-Partner Rollout

Quantium Telstra has announced plans to make Fraud Indicator available to additional partners "shortly", focusing on organisations that interact with customers via digital channels and may be vulnerable to identity fraud. This signals a clear strategic intent: the QT technology is not a bilateral product but a platform designed for multi-organisational deployment.

Target Partners

The most likely expansion targets include:

  • Other major banks (NAB, Westpac, ANZ) — who face identical identity theft threats and would benefit from the same telco intelligence
  • Government agencies — particularly Services Australia, the ATO, and Centrelink, which process millions of identity-verifiable transactions annually
  • Digital services and fintechs — neobanks, payment platforms, and online marketplaces that rely on mobile-based verification
  • Insurance providers — who face fraudulent claims and policy applications

Platform Economics

By positioning Fraud Indicator as a platform rather than a bilateral product, QT creates network effects: each new partner strengthens the underlying ML models with more training data (while preserving individual privacy), which improves detection accuracy for all partners. This is the same principle that makes credit bureaus effective — the system gets smarter as more organisations participate.


Telstra's Broader Scam Protect Ecosystem

Cleaner Pipes Program

The QT partnership sits within Telstra's broader Cleaner Pipes program — a comprehensive initiative to protect customers from scams, malware, and malicious activity across the network. Key components include:

  • Scam Protect: Integrated directly into Telstra's network, available to all retail and wholesale services (including Boost Mobile, Belong, and others) with compatible devices. Customers don't need to activate it — it operates automatically
  • Spam Guard: Real-time identification and blocking of suspected scam calls
  • Content filtering: Network-level blocking of known malicious domains and URLs

The Multi-Layer Defence

Telstra's approach demonstrates a defense-in-depth strategy for scam prevention:

  1. Network layer: Cleaner Pipes blocks malicious traffic at the infrastructure level
  2. Device layer: Scam Protect and Spam Guard operate on compatible devices
  3. Application layer: Scam Indicator and Fraud Indicator integrate with partner applications (banks, government services)
  4. Human layer: Customer education and awareness campaigns

This multi-layered approach recognises that no single control can address the full spectrum of scam and fraud threats.


Regulatory and Industry Context

eIDAS 2.0 and European Parallels

While Australia's approach is still maturing, the European Union's eIDAS 2.0 regulation (implemented in 2024) mandates that all member states provide digital identity wallets with built-in fraud detection capabilities. QT's model — using telco intelligence as a fraud signal — aligns with this direction, suggesting that telco-based identity verification may become a global standard rather than an Australian innovation.

The ePayments Code

Australia's Scheme Rules (formerly the ePayments Code) place increasing obligations on deposit-taking institutions to implement scam prevention measures. The Scam Indicator and Fraud Indicator technologies directly support these obligations by providing banks with tools to detect and prevent fraudulent transactions before they complete.

Government Collaboration

The partnership has received high-level government attention. CommBank CEO Matt Comyn presented the Scam Indicator results to Assistant Treasurer and Minister for Financial Services Stephen Jones during a visit to CBA's South Eveleigh offices — signalling that the government views this as a model for public-private collaboration on fraud prevention.


Competitive Landscape

Alternative Approaches

While QT's telco-intelligence approach is unique in Australia, similar concepts exist elsewhere:

  • UK's MFAS (Match and Fingerprint Authentication Service): Uses device fingerprinting to detect fraudulent account opening
  • US's Plaid: Provides bank-level data verification but lacks the telco dimension
  • India's Aadhaar: Government-run digital identity system with built-in fraud detection

QT's differentiator is the combination of telco data + ML-powered signal extraction + privacy-preserving API sharing — a pattern that could be replicated in other markets with large telco-bank partnerships.

The fundamental insight driving the QT partnership is that identity verification is incomplete without telco intelligence. Mobile numbers are the universal second-factor authentication mechanism across digital services worldwide. When a fraudster provides a mobile number for verification, the telco has unique visibility into whether that number represents a legitimate customer or a fraudster's tool.

No bank, no government agency, and no fintech can replicate this signal. It belongs exclusively to the telco — which is why the QT partnership model (telco + analytics company) is structurally necessary.


Challenges and Criticisms

Data Privacy Concerns

Despite the privacy-preserving design, any system that uses telco data for fraud detection faces legitimate scrutiny:

  • Scope creep: Will the range of signals expand beyond fraud prevention?
  • Data retention: How long are query logs retained, and who has access?
  • Algorithmic transparency: Partners don't see the underlying ML models — how is bias assessed?

QT's response has been to maintain strict purpose-bound access and publish annual impact reports. However, independent oversight mechanisms remain limited.

Coverage Gaps

The system only works for customers with Telstra mobile or landline services. Customers using competing telcos (Optus, TPG, etc.) are not covered by the current integration. This creates a partial protection gap — fraudsters can simply switch to a non-partner telco to avoid detection.

False Positives and Customer Experience

Aggressive fraud detection can create friction for legitimate customers. A freshly provisioned SIM used by a real customer (e.g., someone who just switched providers) may trigger false positives, leading to delayed account openings or additional verification steps that frustrate genuine users.


What This Means for the Industry

The Telco Intelligence Paradigm

The QT partnership demonstrates that telcos hold a critical piece of the identity puzzle that financial institutions and digital services have been missing. As more organisations recognise this, we can expect:

  1. Increased telco-bank partnerships — other telcos (Optus, TPG) will likely develop similar products
  2. Cross-telco interoperability — a future where Fraud Indicator queries work across all major Australian telcos
  3. Regulatory mandates — government may eventually require telco-based identity verification for high-value transactions

The Future of Fraud Detection

The Scam Indicator → Fraud Indicator progression suggests a clear trajectory: from reactive scam detection (during active fraud) to proactive identity verification (before fraud occurs). The next logical step is continuous identity monitoring — real-time assessment of customer identity risk throughout the relationship, not just at onboarding.

Implications for Other Sectors

While the current partnership focuses on banking, the architecture applies equally to:

  • Insurance: Preventing fraudulent claims and policy applications
  • Government services: Verifying identity for Centrelink, tax, and social services
  • Healthcare: Preventing medical identity theft
  • E-commerce: Reducing account takeover fraud and fake reviews

Conclusion

The Telstra-Quantium partnership represents more than a technical innovation — it's a blueprint for cross-industry collaboration in an era where fraud transcends organisational boundaries. By transforming telco intelligence into a privacy-preserving fraud signal, Quantium Telstra has created something that neither company could have built alone: a shared defence against identity theft and financial fraud that protects millions of Australians without compromising individual privacy.

The numbers speak for themselves: 70% reduction in customer scam losses at CommBank, 25%+ increase in identity theft detection, and billions protected from fraudsters who previously operated with impunity.

As fraud becomes more sophisticated — driven by AI-powered social engineering, deepfake voice cloning, and automated account takeover tools — the need for cross-industry intelligence sharing will only grow. The Quantium Telstra model demonstrates that the solution isn't more data silos or stricter individual controls; it's smarter collaboration between organisations that already hold complementary pieces of the puzzle.

For Australian enterprises in any sector that processes customer identities: the question is no longer whether telco intelligence should be part of your fraud detection strategy. It's whether you can afford not to adopt it.


References and Further Reading

  • Telstra and CommBank expand collaboration to increase fraud detection rates (CommBank, Feb 2025): https://www.commbank.com.au/articles/newsroom/2025/02/cba-telstra-fraud-indicator.html
  • Scam Indicator — How It Works (Telstra Exchange): https://www.telstra.com.au/exchange/cyber-security-and-safety/cyber-safety/telstra-scam-indicator---how-it-works
  • Telstra and CommBank expand collab with anti-fraud tech (ARN, Feb 2025): https://www.arnnet.com.au/article/3820325/telstra-and-commbank-expand-collab-with-anti-fraud-tech.html
  • CBA, Telstra partner to protect customers from phone scams (TelecomDrive, Mar 2023): https://telecomdrive.com/cba-telstra-partner-to-protect-customers-from-phone-scams/
  • Quantium Telstra, Commbank launch 'fraud indicator' to lift identity theft detection 25% (Mi3, Feb 2025): https://www.mi-3.com.au/10-02-2025/quantium-telstra-commbank-launch-fraud-indicator-lift-identity-theft-detection-25
  • AI system shields Australians from scams without them knowing (AFR, May 2025): https://www.afr.com/technology/ai-system-shields-australians-from-scams-without-them-knowing-20250515-p5lzko
  • ACCC Scamwatch statistics: https://scamwatch.gov.au/
  • ABS Identity Theft Statistics FY23: https://www.abs.gov.au/

Feature image: Giovanni Domenico Tiepolo / Public domain / Wikimedia Commons.