Qualys today introduced Agent Val, an agentic AI-led workflow inside Qualys Enterprise TruRisk Management (ETM, Qualys's unified platform for managing cyber risk across an organization). It takes on one of the most persistent challenges in vulnerability management: proving what is actually exploitable in your specific environment, right now.
Tami Casey at Qualys described this as a pivotal moment: "Today we unveil Agent Val, the industry's first AI agent for safe exploit validation and autonomous remediation."
The Problem Agent Val Solves
In 2025, more than 48,000 CVEs (Common Vulnerabilities and Exposures, the industry-standard list of publicly known software security flaws) were published. Only a small fraction will ever be remotely exploitable in any given organization's environment. Yet security teams still spend most of their time chasing theoretical risks that may never materialize.
The gap between detection and decision keeps widening. Mandiant's 2024 data showed an average exploitation timeline of -1 day, meaning cyberattacks now routinely happen before a patch even exists. Traditional tools are good at generating exposure scores, but none can actually prove which risks are real and attackable in your specific environment.
How Agent Val Works: A Four-Step Closed Loop
1. Prioritize the Right Less-Than-1%
Agent Val works as a risk-aware selection engine, combining TruRisk scores (Qualys's proprietary risk scoring system), threat intelligence from TruLens, CISA KEV (the U.S. Cybersecurity and Infrastructure Security Agency's Known Exploited Vulnerabilities catalog, a list of vulnerabilities confirmed to be actively used in real attacks), and asset exposure profiles. From tens of thousands of CVEs, it narrows focus to fewer than 10 findings that are both actively weaponized and relevant to your highest-value assets.
2. Prove Exploitability Safely in Production
Using TruConfirm, Agent Val tests whether a vulnerability is actually exploitable. It uses the attacker's real entry path but replaces any destructive payload with a safe verification action. No data is exfiltrated, nothing is written to disk, and there is no production disruption. Just proof. For example, EternalBlue (a well-known Windows exploit) is only flagged as exploitable if SMBv1 (an older, insecure file-sharing protocol) is actually enabled and exposed. Log4Shell (a critical Java vulnerability) is only confirmed if the vulnerable execution path is active in that specific runtime environment.
3. Reduce Risk Intelligently
Once exploitability is confirmed, Agent Val selects the best remediation path: patch, compensating control (a security measure that reduces risk without fixing the underlying vulnerability), or mitigate now and patch later. It uses an AI-powered Patch Reliability Score built from intelligence across 140+ million deployed patches to estimate whether a patch could break something in your environment. For high-risk production systems, it can also apply virtual patches, WAF (Web Application Firewall) rules, network containment, or host isolation while the final patch waits for a safer maintenance window.
4. Revalidate Closure
After remediation, Agent Val reruns TruConfirm against the same asset using the same exploit chain to confirm the attacker path is actually gone. The finding is marked "Exploit Ruled Out" only when re-tested and confirmed closed, not simply when a ticket is marked resolved. This closes a critical gap in CTEM (Continuous Threat Exposure Management, an industry framework for continuously identifying, prioritizing, and resolving security exposures).
How It Fits Into the Broader Platform
Agent Val sits inside Qualys ETM, which brings together asset and identity inventory, risk signals from both Qualys and third-party tools like Microsoft Defender, Wiz, and CrowdStrike, and automated remediation workflows through ITSM (IT Service Management) integrations like ServiceNow and Jira, the platforms organizations use to track and manage IT work. Agent Val connects these tools into a single risk-informed workflow so security, IT operations, and compliance teams all work from the same evidence.
The Business Impact
- 95% reduction in remediation noise - from 62.5 million raw findings down to 150,000 confirmed exploitable exposures
- 60-70% lower MTTR (Mean Time to Remediation) - the average time to fix a confirmed vulnerability, from detection to verified closure
- 10x faster cross-team execution - exploit evidence reduces handoff friction between security, IT, and operations
- Board-ready reporting - CISOs (Chief Information Security Officers) can present hard data on confirmed risk reduction, not just patching activity
Analyst Take
Security has long been treated as an afterthought in the boardroom. It gets labeled a cost center, and the complexity of the threat landscape makes it genuinely hard for business leaders to engage with it. That is starting to change. Boards know a breach is not a question of if but when, and security investment is simply the cost of operating in a connected world.
The real challenge for security leaders is communication. They need to translate a constantly shifting threat landscape into language that resonates with executives and boards. Not vulnerability counts and patch rates, but business risk, financial exposure, and evidence that defenses are actually working.
This is where Qualys TruRisk, and Agent Val specifically, addresses a real gap. Moving from theoretical risk scores to confirmed, validated exposure gives security leaders something concrete to bring to the table. Not "we patched 10,000 vulnerabilities" but "we confirmed and closed 150 real attack paths." That is a more credible conversation.
The question for organizations looking at this is not whether prioritization matters. It clearly does. The real question is whether running AI-driven live exploit validation in production is a step they are ready for operationally and culturally. For mature security teams, this could be a real force multiplier. For others, getting visibility and risk prioritization right comes first, and TruRisk is a solid place to start.
Why This Matters
The industry has long automated assumptions, patching faster based on severity scores without checking whether a vulnerability is actually exploitable in a specific environment. Agent Val shifts that model to proven risk reduction with a closed loop: validate, mitigate, revalidate.
Read the full Qualys blog post: Meet Agent Val: Closing the Validation Gap in Exposure Management at Machine Speed with Agentic AI
