eSentire has released Atlas Preempt for continuous pentesting operations using AI. Both offensive and defensive SecOps teams now have a tool to identify exposed systems at risk of exploitation. With agentic AI continuously probing and looking for vulnerabilities, this gives defenders a chance to remediate before a breach.
The company says that this is a critical pillar of its Atlas Platform, and the results are fed into eSentire’s 24/7 Managed Detection and Response (MDR) solution.
Dustin Hillard, Chief Product and Technology Officer, eSentire, said, “The economic advantage has always favored attackers, and frontier-AI models widen it. They now operate at machine speed and scale, while quarterly assessments and severity scores were built for a threat that moved at human pace.
“Atlas Preempt closes that gap, giving defenders a continuous attacker’s-eye view of their own environment and turning vulnerability disclosures into validated, prioritized risk in hours, not weeks, with the human oversight boards, regulators, and insurers require.”
Agentic AI has changed the speed at which attackers can compromise systems. In November 2025, researchers from Anthropic documented the first confirmed AI-orchestrated espionage campaign. It showed how autonomous agents executed 80% to 90% of offensive steps. Many of those steps, such as reconnaissance, vulnerability scanning, credential harvesting and privilege escalation, are labour-intensive.
When used by threat actors, autonomous AI removes those bottlenecks. It can chain multiple attack stages, enabling it to move faster than defenders can identify. A review by Sprinto of incidents across 2025 and 2026 shows how effectively AI can locate and exploit misconfigurations and vulnerabilities. For the attackers, it means no need to learn coding. Instead, they just use natural language instructions to bypass defences.
The tools used by defenders are reactive by nature. Despite attempts to move defences into a more reactive mode, defenders are often split between red and blue teams. Red attacks and blue defends and fixes. Shifting to a purple team approach is about collaborative security, rather than adversarial security.
The SANS Institute has an exceptionally good paper on the importance of Purple Teaming and how it improves defence. In an AI era, this is something that organisations must look to adopt.
Atlas Preempt acts as an AI-driven attacker but delivers the output to SecOps teams. Among its capabilities are:
How Atlas Preempt orchestrates all of this is important. For SecOps teams, it gives them a view into how attacks are evolving. It also gives them an opportunity to understand other weaknesses in their defences. That latter point is critical. Organisations are still focused on perimeter defence and not doing enough to disrupt attackers once they are inside the network.
Atlas Preempt will also allow organisations to move to a risk-driven approach. It will show where the greatest and most obvious vulnerabilities are. That creates a priority patch plan. It moves organisations away from trying to patch everything to laser-focused patching. It also removes the risk of a low-priority exploit being ignored and leaving access open to an attacker.
The announcement came with an example of how an unnamed healthcare provider used eSentire’s tools to protect its electronic medical record (EMR) system. It caught a critical exposure in a staging API host that had been missed in previous testing. The miss was caused by the complexity of the system, which has over 300 subdomains that the AI testing found.
It testing, it discovered that authorisation controls had been inadvertently disabled across all 106 endpoints. That single mistake left 20 million patient records exposed. Among the data that could be access was social security numbers, prescriptions, dates of birth, and addresses.
eSentire says that it “validated and delivered these findings to the customer on the same day, alongside a five-step prioritized remediation plan, enabling immediate web application firewall containment.
“The end result: The organization remained ahead of its 60-day HIPAA breach assessment timeline and this exposure was identified, remediated and closed before an attacker had the opportunity to take advantage of it.”
Security vendors have spent the last few years spreading fear and doom around the risk from AI. That has now come to pass, yet many have only added AI to help with analysts in the SOC. While they were focused on that, attackers have successfully built and deployed AI-based solutions.
We are now seeing a phase of catch-up from security tools vendors. They are adopting a Risk Operation Centre approach to vulnerabilities and patching.
eSentire is not alone in the autonomous pentesting market. AttackIQ and SafeBreach both have platforms that focus on breach and attack simulation (BAS). What is different here is the way that eSentire has integrated Atlas Preempt into its MDR, the Atlas Platform. This means that it becomes the key element required for Purple Teaming.
The customer story from healthcare is interesting. What will be more interesting will be the publication of case studies that look in more detail at what Atlas Preempt uncovers. That will be the moment when customers will see real ROI. At the moment, the story is still about reducing exploit time and being continuous. Customers want a number that applies to their bottom line.
How will this play out in cyber insurance? eSentire has relationships with cyber insurers. It would be a shock if they were not looking at how Atlas Preempt reduces risk and exposure. After all, they don’t want to pay out if they can avoid it.
The questions now are, how fast will eSentire’s customers deploy Atlas Preempt? When will we see more case studies on its impact? Where are the tutorials on how to best use it? What additional integrations will eSentire deliver?
Customers who want to understand more can sign up for complimentary 30-day access to the Atlas Platform, including vulnerability assessment and penetration testing.
Enterprise Times – technology for your enterprise today and tomorrow.
© 2015 – 2023 Synonym Ltd
About | Meet the Team| Terms | Privacy | Copyright | Cookies | Contact | Newsletter
Cookie Settings