CERT-In Warns Payment Fraud Is Moving Through Altered Records and AI-Driven Tools
Key Facts
- CERT-In issued critical advisory CIAD-2026-0062 on November 20, 2026, warning of unauthorised financial transactions at BFSI service providers, government-linked payment technology providers and vendor or partner payout platforms.
- Attackers altered business records or approval relationships and then used legitimate payment processes; CERT-In did not disclose amounts, transaction counts or affected organisations.
- The advisory flags AI-driven agentic offensive frameworks used in attacks on enterprises, technology companies, service providers, government bodies and academic institutions.
- CERT-In recommended independent reconciliation of payment records, independent verification of high-risk changes, separation of client environments and restrictions on administrative access.
- CERT-In reiterated its six-hour incident-reporting requirement and the National Cyber Crime Reporting Portal and 1930 helpline for unauthorised transactions.
CERT-In issued a critical advisory, CIAD-2026-0062, on November 20, 2026, warning that unauthorised financial transactions are occurring at service providers for banks and financial-services firms, at technology providers supporting government-linked payments, and at platforms that handle vendor or partner payouts.
The mechanism is not a direct break of payment APIs. Attackers altered business records or approval relationships, then used legitimate payment processes to move funds, CERT-In said. A transaction can pass routine application and approval checks even when the records used to initiate it were changed without authorisation. Gaps in independent verification, reconciliation and review of record changes make the fraud harder to detect.
CERT-In did not disclose the amounts involved, the number of transactions or the organisations affected.
The advisory also says CERT-In’s threat-hunting teams identified AI-driven agentic offensive frameworks in attacks on enterprises, technology companies, service providers, government organisations and academic and research institutions. The frameworks can automate reconnaissance, prepare or modify exploits for known vulnerabilities, identify privileged access and run multiple tasks in parallel, and can adjust scripts and requests when they encounter errors or blocks. They may use valid credentials, legitimate administrative functions and trusted relationships to gain further access. CERT-In said their use does not mean an AI agent can independently decide to attack an organisation.
CERT-In recommended that organisations reconcile payment instructions and critical business records against independent authoritative sources that cannot be altered through the same compromised provider environment. It also called for independent verification of high-risk changes to payment instructions, beneficiaries and approval arrangements, separation of client environments, restrictions on administrative and control-plane privileges, and monitoring of record changes alongside privileged activity and the transactions that follow.
The agency reiterated its six-hour incident-reporting requirement and said cases involving unauthorised financial transactions should also be reported through the National Cyber Crime Reporting Portal or the 1930 helpline.
The advisory follows a November 4 CERT-In warning about the AI-driven CRAFTY FALCON campaign against financial firms, which named tools including PentAGI, Cairn and CyberStrikeAI, and a Reserve Bank of India warning to banks about rising cyberattacks and AI-related risks. CIAD-2026-0062 does not name a CVE or malware family for the payment incidents it describes.
Sources & References
- CERT-In advisory CIAD-2026-0062, Abuse of Business Payment Workflows and Use of AI-Driven Agentic Offensive Frameworks across Sectors, issued November 20, 2026. https://www.cert-in.org.in/s2cMainServlet?pageid=PUBVLNOTES02&VLCODE=CIAD-2026-0062
SecurityXP delivers cybersecurity news, vulnerability analysis, data breach reports, and threat intelligence.
Security Digest
The most important cybersecurity stories, CVE alerts, and threat intelligence — one concise email. Unsubscribe anytime.
Related Articles
Chinese-Nexus Hackers Used an Exposed Cairn AI Instance to Find and Exploit an Apparent Zero-Day
Amazon Threat Intelligence says Chinese-nexus hackers used Cairn, an open-source autonomous AI penetration-testing framework, to discover an apparent…
FBI Criminal Database Breached Again, Second Hack in Months Targets US Crime Records
The FBI's criminal database has been hacked for the second time in months, according to reports from India Today and Hindustan Times.
n0n Publishes 16,000 JPMorgan Reserve Card Emails After Vendor Breach
Roughly 16,000 customer email addresses have been published in a dataset apparently linked to JPMorgan Chase's invitation-only Reserve card program…
Bell Canada Confirms "Agentic" Threat Actor Breached Systems and Stole Enterprise Customer Data
Bell Canada says an "agentic" threat actor accessed a pre-production Maximo environment on November 17 and exfiltrated enterprise customer data, but has not attributed the attack.