AI security tools gain agentic pentesting, attack surface management, and data flow protection

Five security vendors released new AI-powered tools on July 28, 2026, targeting gaps in continuous penetration testing, attack surface management, data security, and threat intelligence. Bugcrowd, Prescient Security, Cyberhaven, Intel 471, and SpecterOps each announced products that use AI agents to automate or augment security workflows.

Bugcrowd introduced Savant Pathseeker, the first product in its Agentic Offensive Testing line. The tool uses purpose-built AI agents to continuously test external web applications and APIs, validating exploits and providing reproducible proof of exploitability. Findings come with audit-ready reporting, and users retain on-demand access to Bugcrowd's human pentester community for complex work such as business logic flaws and zero-days.

Prescient Security and Cyberhaven expand AI-assisted coverage

Prescient Security announced a series of updates to Cait, its continuous AI-assisted penetration testing service. The updates, rolling out through summer 2026, add attack surface management, new asset testing types, and expanded environment support beyond web applications. Cait acts as an autonomous, context-aware pentester that explores applications before attacking them.

Cyberhaven launched Cyberhaven Flow, an AI-native data security platform designed to protect data across human and AI workflows. Flow connects data lineage, identity, and behavior to track data as it is created, copied, fragmented, and shared across endpoints, browsers, and cloud environments. The platform aims to adapt protection to the changing context of work in the AI era.

Intel 471 and SpecterOps add AI agents for threat intelligence and identity security

Intel 471 expanded its Verity471 platform with two new AI capabilities: MCP471 and Agent471. These tools aim to make threat intelligence more accessible by helping security teams pinpoint relevant threats and pair them with internal telemetry. The platform turns adversary tradecraft into pre-attack intelligence and proactive threat hunting.

SpecterOps announced new capabilities for BloodHound Enterprise, adding support for Amazon Web Services and Microsoft Entra Agent ID. The company also introduced BloodHound Hunter, a purpose-built AI agent interface that brings adversary intelligence into AI-assisted security workflows. The tools help defenders understand how adversaries traverse hybrid environments and proactively eliminate attack paths.

Market context and what comes next

The announcements reflect a broader shift in the security industry as defenders race to match attackers who are already using AI to probe continuously. Traditional periodic manual pentests leave assets exposed for months, while high-volume scanners generate noise without confirming exploitability. Agentic pentesting solutions aim to combine AI speed and scale with human expertise.

Bugcrowd stated it does not use customer or researcher data to train the models powering Savant Pathseeker. The product runs on frontier AI models built on a proprietary layer of skills developed by Bugcrowd's in-house practitioners. Future iterations will involve Bugcrowd's Hacker Advisory Board to develop an incentive framework for researcher contributions.