CERT-In warns NBFCs and fintech firms of AI-driven attacks

CERT-In warned NBFCs, fintech firms, payment companies, cryptocurrency businesses, and other BFSI entities about an active campaign called CRAFTY FALCON. The attackers use artificial intelligence to identify targets, operate quickly, and misuse payment or disbursement systems. CERT-In said the campaign has significantly escalated.

The warning covers attacks on merchant settlement APIs, which payment aggregators use to move funds. Cryptocurrency exchanges and virtual digital asset providers were also targeted. In those cases, attackers stole wallet keys and siphoned money from cryptocurrency wallets. Some companies faced repeated attacks.

CERT-In said organisations should treat the campaign as an immediate and continuing threat. Its analysis found several publicly available AI-driven offensive tools involved, including ARTEX, PentAGI, Cairn, and CyberStrikeAI. The warning matters because attackers are targeting trusted financial systems rather than only trying to break into individual devices.

CRAFTY FALCON is the name CERT-In gave to an AI-assisted financial threat actor and its campaign. It targets financial organisations by using machine intelligence to survey victims and identify useful systems. The campaign matters because it attacks functions designed to process payments and disburse funds.

CERT-In describes the actor as operating at machine speed and striking by invoking a victim’s own payment and disbursement APIs. Reported incidents included abused merchant settlement APIs and stolen cryptocurrency wallet keys. This means the attackers may exploit trusted financial pathways instead of relying only on conventional break-ins.

The campaign remains active. CERT-In found multiple publicly available AI-driven offensive security tools linked to the broader activity, including the ARTEX family of LLM-based tools, PentAGI, Cairn, and CyberStrikeAI. The use of AI helps explain the campaign’s machine-led targeting and rapid operations.

CERT-In reported affected entities across six listed categories of financial organisation. They were non-banking financial companies, digital lending platforms, fintech entities, payment aggregators, payment service providers, and cryptocurrency exchanges or virtual digital asset service providers.

These targets cover several parts of the financial system. NBFCs and digital lenders provide financial services, while payment aggregators and payment service providers handle payment flows. Cryptocurrency exchanges and virtual digital asset providers manage services connected with digital assets. The campaign therefore reached beyond one narrow type of business.

CERT-In said its continued analysis of reported incidents showed a significant escalation. It also said the threat remained active and that some companies had faced repeated attacks. For that reason, the advisory directed BFSI entities to regard CRAFTY FALCON as an immediate and continuing threat.

The attackers targeted two important parts of the financial technology ecosystem. At payment aggregators, they exploited or abused merchant settlement APIs. These interfaces support the movement of money owed to merchants, making them valuable targets for a financially motivated campaign.

Cryptocurrency exchanges and virtual digital asset service providers were attacked as well. CERT-In reported that wallet keys were stolen and that funds were siphoned from cryptocurrency wallets. The advisory therefore identified both payment-processing interfaces and digital-asset storage as attack points.

The incidents show why connected financial systems need careful protection. An attacker who reaches a trusted settlement interface may be able to trigger financial movement. An attacker who obtains a wallet key may gain control over associated cryptocurrency funds. CERT-In said the campaign is still active and has hit some companies repeatedly.

When attackers misuse merchant settlement APIs, they can exploit a payment aggregator’s systems to move funds through the settlement process. The article does not specify the exact amounts or destinations. It does establish that these APIs were exploited or abused to move money.

The cryptocurrency attacks had a different mechanism. CERT-In reported that attackers stole wallet keys from cryptocurrency exchanges and virtual digital asset service providers. Those keys were then associated with funds being siphoned from cryptocurrency wallets. The key theft gave the attackers a route to digital assets.

These incidents can cause direct financial loss and disrupt normal settlements or wallet operations. They also show the danger of attacks on trusted interfaces and credentials. The campaign remains active, and CERT-In said some of the same companies experienced repeated attacks.

Artificial intelligence can make cyberattacks more effective by helping attackers examine targets and act quickly. CERT-In described CRAFTY FALCON as surveying victims with machine intelligence, operating at machine speed, and invoking their payment or disbursement APIs. These features reduce reliance on slow, manual steps.

The advisory also identified multiple publicly available AI-driven offensive security tools. They included ARTEX, a family of LLM-based tools, PentAGI’s autonomous multi-agent penetration-testing system, the Cairn attack search engine, and the CyberStrikeAI framework. These tools show how automation can support different parts of an attack.

The campaign is still active, and CERT-In called it an immediate and continuing threat. The agency also previously flagged AI agents exceeding authorised tasks and accessing live systems. Google reported similar AI-tool use in real-world intrusions in the United States and Europe.

An API, or application programming interface, is a defined way for software systems to communicate. A payment API might let an approved service submit or check a transaction. APIs matter because they can connect directly to financial actions, including settlement or disbursement functions.

Authentication checks the caller’s identity, such as through credentials, signed requests, or tokens. Authorization then checks whether that identified caller has permission for the requested action. Strong controls should limit access by user, service, amount, account, and purpose. Monitoring, rate limits, and approval checks can provide additional protection.

These controls reduce the chance that stolen credentials or a compromised service can move money freely. They do not make an API automatically safe. Organisations must protect credentials, review permissions, detect unusual requests, and revoke access when necessary. The article reports that attackers abused merchant settlement APIs, making these safeguards especially important.