Offensive security solutions firm HackerOne Inc. today announced the expansion of its artificial intelligence-driven offensive security platform.

The updates includes the evolution of its Hai AI system into a coordinated team of agentic AI agents and the general availability of its AI-native code security product, HackerOne Code. They’re aimed at redefining continuous threat exposure management by accelerating how enterprises identify, validate and remediate vulnerabilities while reducing manual effort and time to resolution.

HackerOne’s Hai AI agent, which originally launched in February 2024, is evolving from a copilot-style offering to a team of AI agents that continuously analyze and contextualize findings to help organizations prioritize, validate and remediate risk faster.

The new version of Hai is launching with four agents: Priority Escalation Agent uncovers critical risks without delay, Deduplication Agent eliminates duplicates and reduces noise, Report Assistant Agent ensures complete, consistent reports and Insight Agent surfaces historical context to accelerate validation.

In testing before its official launch, the new Hai agents were found to deliver time savings, with users saving up to 40 hours or more every month. “Hai cut our validation time from 20 minutes to just five,” said Connor Knabe, application security architect at Veterans United Home Loans Inc. “By replacing manual steps with clear context, we validate faster, clarify impact and stay aligned.”

Alongside the upgrade to Hai, HackerOne announced the general availability of HackerOne Code to solve the challenge of discovering and eliminating vulnerabilities before applications are deployed.

HackerOne Code is a code security solution that works like a developer and thinks like a security researcher — scaling vulnerability discovery with AI and human oversight and has been built for the AI development era.

The company also today previewed the next milestone in its agentic roadmap: Agentic Pentesting.

Agentic Pentesting is aimed at continuously provig exploitability at an AI-driven scale while keeping human ingenuity at the core. The service will extend exposure management into adversarial validation to deliver real proof of exploitation so organizations can prioritize and remediate with greater confidence.

“Hai has been central to our vision for AI-powered offensive security and today marks the next stage in its evolution,” said Nidhi Aggarwal, chief product officer at HackerOne. “Powered by the insights drawn from over a decade of offensive security expertise, Hai’s new agents and the introduction of Agentic Pentesting extend its capabilities from validation to proof of exploitability, helping organizations continuously reduce exposure and accelerate remediation at scale.”

HackerOne is a venture capital-backed startup that has raised $159 million across six rounds, including a round of $36.4 million in September 2019. Investors include Valor Equity Partners LP, Benchmark Capital Partners, New Enterprise Associates Inc., Dragoneer Investment Group and EQT Ventures.

Photo: HackerOne