AI-Assisted Autonomous Cyber Attacks Targeting Digital Banking and Financial Services

Published on 17-Nov-2026 14:30:00

Executive Summary

BGD e-GOV CIRT's Cyber Threat Intelligence (CTI) Unit is issuing this advisory to alert banks, financial institutions, fintech organizations, mobile financial service (MFS) providers, payment service providers and other organizations operating Bangladesh's financial ecosystem about the increasing threat of AI-assisted autonomous cyber-attacks.

Recent threat intelligence has identified sophisticated cyber operations in which attackers used AI-driven penetration-testing frameworks and autonomous agents to automate reconnaissance, vulnerability identification, exploit development, credential harvesting, privilege escalation, lateral movement and data exfiltration.

The investigation documented a full-chain compromise of Tanner Servicios Financieros, a financial-services organization in Chile. Attackers reportedly gained remote code execution in a Kubernetes-hosted application environment and subsequently moved through enterprise identity systems, Microsoft SQL Server environments, middleware and cloud infrastructure. The investigation reported access to more than 15.6 million financial records.

Separate evidence documented attack preparation against Dengoo.kz, an online micro-lending platform in Kazakhstan, involving a malicious WordPress theme and a custom PHP web shell. Successful web shell execution against that organization was not confirmed in the available evidence.

The investigation also identified infrastructure and activity associated with financial-sector reconnaissance and attack operations in other countries. Some additional fintech organizations were identified in attacker-side configuration files, and probing activity was independently reported; however, their compromise was not established.

The observed attack chain demonstrates how autonomous AI-assisted tooling can coordinate multiple conventional attack techniques and accelerate intrusion operations.

Assessment for Bangladesh: The techniques are relevant to Bangladesh's financial ecosystem because banks and financial-service organizations operate interconnected internet-facing applications, APIs, databases, identity infrastructure, cloud platforms and third-party services. Organizations should assess these environments proactively. This advisory does not claim that any Bangladeshi institution has been compromised in the reported incidents.

Why AI-Assisted Attacks Are a Concern

Traditional intrusion operations often require an attacker to manually perform reconnaissance, interpret results, select attack techniques and develop custom scripts.

AI-assisted agents can automate or coordinate several of these tasks. The investigated operation reportedly involved:

·   Automated discovery of internet-facing services.

·   Identification of exposed administrative interfaces.

·   Generation of customized scripts and payloads.

·   Automated testing of application weaknesses.

·   Credential discovery and internal network enumeration.

·   Selection of privilege-escalation and lateral-movement paths.

·   Collection and processing of sensitive information.

·   Adaptation of scripts based on target responses.

The principal concern is not simply the use of AI. It is the ability to automate multiple stages of an intrusion and conduct them at greater speed and scale.