Protectt.ai launches AI security platform for banks

Protectt.ai has launched an AI for Security Platform for banks, non-banking financial companies and insurers. The product is designed to test web and mobile applications during software release cycles.

The Mumbai-based security company said the platform uses autonomous AI agents to simulate cyberattacks on applications before they reach customers. The system is intended to bring red-team testing into routine software updates rather than limit checks to periodic assessments.

Financial institutions are releasing software more frequently, with some updating applications weekly or daily. That pace can create gaps between formal security reviews, during which new weaknesses may enter production systems.

The platform continuously probes applications for vulnerabilities and assesses whether separate weaknesses could be chained together and exploited. This is intended to help security teams test each new release instead of waiting for the next scheduled review.

"Security cannot be a quarterly exercise when financial applications are changing every day," said Manish Mimani, Founder and Chief Executive Officer of Protectt.ai.

"With autonomous AI agents, organisations can put their applications under continuous attack to do more than just find vulnerabilities. They can see how a real attacker could discover, connect and exploit them. The goal is to make red-team testing part of every release, so security teams can identify the path to an attack before an attacker does," said Mimani.

Platform tools

The platform combines three products. RedPilot is an autonomous red-teaming tool designed to find and connect security weaknesses in a way that mirrors attacker behaviour.

WebScan is designed to test web applications, networks, infrastructure and application programming interfaces against more than 30 attack scenarios. AppScan, Protectt.ai's mobile application security product, has also been updated with AI and runs more than 140 automated security tests on each application build, according to the company.

The platform is aimed at regulated organisations that may need tighter control over where testing takes place and where security data is stored. The AI agents can run within a bank's own environment, including on-premises, private cloud and hybrid setups.

According to Protectt.ai, application code and security findings are not sent to external AI services. Testing remains authorised and controlled by an organisation's own security team.

Regulatory setting

The launch comes as financial institutions face regulatory requirements for security testing alongside internal technology governance demands. In India, the Reserve Bank of India's Master Direction on IT Governance, Risk, Controls and Assurance Practises requires vulnerability assessments for critical systems at least twice a year and penetration testing at least once a year, as well as after major changes.

Protectt.ai said the platform is designed to support ongoing testing alongside those requirements rather than replace the independent assessments required by regulators or boards. It is also not intended to replace human security specialists.

Tasks such as social engineering, physical security reviews and some business-logic issues still require human judgment, the company said. The AI agents are instead intended to automate repeatable testing tasks that would otherwise be difficult to run with every software release.

The platform also produces reports with evidence of identified security issues and guidance for remediation, according to the company. These are intended to help security teams assess risk, fix vulnerabilities and track changes across successive releases.

Protectt.ai focuses on mobile application and AI security for banks, financial institutions and other organisations. Backed by Bessemer Venture Partners, the company said it protects more than 570 mobile apps, covers more than 1 billion unique devices and has prevented more than 8 billion threats.

It operates from offices in India, the UAE and the US, and said it safeguards more than 100 billion sessions.