Dominic Jainy is a distinguished IT veteran whose expertise bridges the complex worlds of artificial intelligence and cybersecurity. With a career dedicated to securing enterprise infrastructures, he has become a go-to expert for understanding how advanced machine learning models can both expose and defend against high-stakes vulnerabilities. In this conversation, we explore the recent wave of critical security disclosures affecting Ivanti’s most prominent platforms—Endpoint Manager Mobile, Neurons for ITSM, and Sentry—and what these findings reveal about the modern state of automated threat detection and incident response.
The recent disclosures regarding Ivanti Neurons for ITSM have highlighted several critical flaws, specifically those with a CVSS score of 9.8 and 9.9. Could you elaborate on the technical nature of these deserialization vulnerabilities and why they represent such a severe threat to enterprise servers?
The situation with Ivanti Neurons for ITSM is particularly alarming because of how deeply these vulnerabilities penetrate the trust model of a server. When we look at CVE-2026-12744 and CVE-2026-12745, we are dealing with the deserialization of untrusted data, which is essentially a “skeleton key” for unauthenticated attackers. By sending a specially crafted bit of data, an attacker can force the server to execute arbitrary code without ever needing a valid username or password. This is why they carry that terrifying 9.8 severity rating; it is a direct line into the heart of the corporate infrastructure. Furthermore, the 9.9 rated issues, like CVE-2026-12645, focus on missing authorization, meaning even an authenticated user with low privileges could pivot and run code with administrative authority.
Ivanti explicitly credited advanced large language models for uncovering several of these ITSM flaws. How does this shift in using AI-assisted discovery change the way we view vulnerability management and the speed at which companies must now respond?
This is a landmark moment in our industry because it’s one of the first times we’ve seen a major vendor formally credit LLMs in a security advisory for finding critical bugs. It indicates that Ivanti has successfully integrated AI into its product security and engineering workflows to look for flaws that human eyes might have missed for years. For vulnerability management teams, this means the “arms race” has entered a new phase where the speed of discovery is accelerating exponentially. If companies can find their own zero-days using AI, you can bet that sophisticated threat actors are using the same models to hunt for exploits. This creates an environment where the window between discovery and exploitation is shrinking, demanding that our response plans become just as automated as the discovery process itself.
Beyond the ITSM platform, the vulnerabilities in Ivanti Sentry and Endpoint Manager Mobile (EPMM) seem to target the very tools meant to secure mobile workforces. What are the specific risks associated with the authentication bypass and privilege escalation flaws found in these products?
The risks here are visceral because EPMM and Sentry are the gatekeepers for an organization’s mobile fleet, and the flaws found there are quite dangerous. For instance, CVE-2026-18851 in EPMM allows a remote authenticated attacker to escalate their privileges to full administrator status, which is a high-severity 8.8 risk. Imagine a disgruntled employee or a compromised user account suddenly gaining the keys to manage every mobile device in the company. Then you have CVE-2026-83527 in Ivanti Sentry, which is an authentication bypass rated at 8.1 that allows unauthenticated administrative access. This vulnerability, responsibly disclosed by researcher btaol of Aquila Sec Lab, effectively removes the lock from the front door, allowing an outsider to walk right into the management console and wreak havoc on the entire mobile ecosystem.
The remediation process for these vulnerabilities seems to differ significantly between cloud and on-premises customers. Could you walk us through the timelines for these patches and what on-premises administrators should be doing right now to protect their environments?
The contrast in remediation timelines really highlights the inherent security advantages of the SaaS model in 2026. Cloud and SaaS versions of Neurons for ITSM were automatically patched across all landscapes back on August 9, 2026, so those customers didn’t have to lift a finger to be protected. On-premises customers, however, have a much heavier burden; those running versions 2025.2 through 2026.1 must manually apply the September 2026 security patches immediately. The pressure is even higher for those waiting on version 2026.2 for on-premises deployments, which isn’t scheduled for release until September 21. Until that patch is live and installed, these administrators are operating in a state of high tension, especially if their instances are internet-exposed, and they must prioritize these updates over almost any other IT task to avoid becoming a headline.
What is your forecast for the security landscape of edge and mobile management infrastructure over the coming months?
My forecast is that we are about to witness a significant surge in targeted attacks against these specific Ivanti platforms as threat actors race to reverse-engineer the patches released on September 8, 2026. Even though Ivanti stated there was no evidence of active exploitation prior to the disclosure, history tells us that edge infrastructure is a favorite target for state-sponsored groups and ransomware operators. I expect to see the first wave of automated scanning for these CVEs within the next 48 to 72 hours, focusing heavily on those unpatched on-premises ITSM instances. Organizations that fail to move to fixed versions like 12.10.0.0 for EPMM or R10.8.2 for Sentry will likely face a breach before the year is out. Ultimately, this disclosure will push more enterprises to abandon on-premises management altogether in favor of cloud environments where the vendor can mitigate these critical 9.9 risks instantly.