July’s enterprise patch cycle delivered a blunt warning: AI is changing vulnerability management from both sides.
Microsoft released its largest Patch Tuesday on record, with security researchers reporting between 570 and 622 vulnerabilities depending on counting methodology. ServiceNow patched a critical remote code execution vulnerability in its AI Platform. SAP’s July Security Patch Day addressed critical issues in NetWeaver Application Server ABAP, SAP Approuter, and SAP Commerce Cloud.
The updates show how enterprise security teams are being pulled into a faster, more complicated patching environment. AI-assisted vulnerability discovery is helping vendors and researchers find more flaws. Attackers are also using AI to accelerate exploit development. At the same time, AI platforms and agentic systems are becoming part of the attack surface.
For ERP leaders, the lesson is not limited to security operations. Modern ERP estates depend on Microsoft infrastructure, SAP applications, ServiceNow workflows, cloud services, identity platforms, integration layers, and AI-enabled automation. Patch management is becoming a business-continuity issue for the systems that run finance, procurement, HR, manufacturing, service, commerce, and supply chain operations.
Analysis
What this means: Patch management is an AI-era capacity problem. Security teams are not only dealing with more vulnerabilities; they are dealing with faster discovery, shorter exploit windows, and more critical platforms sitting inside business workflows. ERP leaders need patch readiness to cover infrastructure, application layers, AI services, and the workflow platforms that connect them.
Microsoft Breaks the Patch Tuesday Record
Microsoft’s July Patch Tuesday reset expectations for enterprise patch volume.
SecurityWeek reported Microsoft patched a record 622 vulnerabilities, including two zero-days that had been exploited in the wild. Help Net Security counted the release as 570+ vulnerabilities, while also highlighting the two exploited flaws and one publicly disclosed issue. The difference reflects counting methodology, especially around Chromium-related fixes and broader ecosystem tracking.
The exploited vulnerabilities were CVE-2026-56155, an elevation-of-privilege flaw affecting Active Directory Federation Services, and CVE-2026-56164, an elevation-of-privilege flaw in Microsoft SharePoint Server. Microsoft also addressed CVE-2026-50661, a publicly disclosed BitLocker security feature bypass vulnerability that had not been observed under active exploitation at the time of reporting.
The Microsoft cycle touched systems many ERP organizations depend on directly or indirectly: Windows Server, SharePoint, identity infrastructure, Office, Microsoft 365, Dynamics environments, Azure-connected workloads, developer tooling, and endpoint estates that support enterprise applications.
The larger story is why the patch count is rising.
Microsoft had already warned customers to expect larger security releases as AI helps defenders find more vulnerabilities faster. In a July 9 Windows blog, the company said AI-assisted discovery means customers will see a higher volume of security updates in each release. Microsoft also said it is updating its Secure Development Lifecycle to account for AI-enabled attack techniques and exploit paths.
Microsoft’s MDASH system, short for Multi-Model Agentic Scanning Harness, is central to that shift. The company describes MDASH as a multi-model agentic scanning system that uses specialized AI agents and multiple models to discover, validate, and help remediate software vulnerabilities across complex codebases.
That creates a new patching reality for customers. Larger releases may indicate better defensive discovery, not necessarily worse software quality. But the operational burden still lands on enterprises that must test, prioritize, deploy, and monitor updates without breaking critical systems.
AI Raises the Tempo
Microsoft’s guidance points to a harder operating model for vulnerability management.
Microsoft now recommends deploying Windows quality updates with less than three days of deferral, deadlines of zero or one day, and a grace period of no more than two days. That is a sharp message for enterprises used to slower patch testing windows.
The challenge is ERP environments often rely on shared infrastructure and connected services that cannot be patched casually. A Windows Server update, SharePoint fix, identity-system hardening step, or endpoint change can affect integrations, approvals, reporting, workflow automation, file exchange, and user access.
The old patching model assumed security teams could triage vulnerability lists at human speed, then work through testing and deployment windows based on severity, exploitability, and operational risk. AI compresses that timeline. Attackers can analyze disclosed fixes faster. Researchers can generate proof-of-concept exploit paths faster. Vendors can uncover larger backlogs of latent vulnerabilities faster.
The result is a widening gap between vulnerability discovery and enterprise remediation capacity.
Analysis
What this means: AI is breaking the old patch calendar. Monthly release cycles still exist, but exploit analysis is moving faster than many enterprise change-control processes. Organizations running ERP on Microsoft infrastructure need a patch model that separates emergency exposure from routine maintenance, especially for identity, collaboration, and server components tied to core business systems.
ServiceNow’s AI Platform Gets a Critical Fix
ServiceNow’s July patch story is narrower but relevant to the enterprise AI conversation.
The company patched CVE-2026-6875, a critical remote code execution vulnerability identified in the ServiceNow AI Platform. The vulnerability carries a CVSS 4.0 score of 9.5 and could allow an unauthenticated user, under certain circumstances, to execute code within the ServiceNow platform.
ServiceNow addressed the issue by deploying a security update to hosted instances and providing relevant updates to self-hosted customers and partners. Vulnerability tracking sources said ServiceNow was not aware of active exploitation against instances at the time of disclosure.
The Canadian Centre for Cyber Security also issued an advisory encouraging users and administrators to review the ServiceNow guidance and apply necessary updates. The advisory listed affected releases across Brazil, Australia, Zurich, and Yokohama versions before specified patched releases.
The technical reporting described the issue as a sandbox escape in the AI Platform layer. That description is important because AI platform features often rely on controlled execution environments, scripts, automation, data access, and workflow context. A sandbox escape weakens the boundary that is supposed to keep execution contained.
For customers, the risk is not only that an AI feature has a bug. AI capabilities are increasingly embedded inside platforms that already orchestrate IT service management, cyber operations, employee workflows, HR service delivery, finance support, procurement processes, and enterprise operations.
AI Layer Part of Attack Surface
ServiceNow’s patch highlights a shift many enterprises are still absorbing.
AI functionality is being built into platforms that route requests, trigger workflows, update records, escalate issues, summarize evidence, and coordinate work across departments. That makes the security model more complicated. Customers need to know where AI services execute code, what data they can access, which identities they use, how sandboxing works, which plugins or skills are enabled, and how quickly cloud and self-hosted deployments receive security updates.
The ServiceNow vulnerability also shows why AI governance cannot stop at prompt safety or model behavior. Traditional application security still applies. So do patching, runtime isolation, least privilege, logging, vulnerability scanning, change management, and incident response.
For ERP leaders, ServiceNow’s role in enterprise workflows creates a direct operational link. A ServiceNow platform weakness can affect ticketing, access requests, change approvals, incident response, security operations, employee service delivery, and support processes connected to ERP environments.
Analysis
What this means: AI platforms need the same patch discipline as core enterprise applications. The risk is not limited to what an AI assistant says; it includes where the AI layer runs, what it can touch, and whether its execution boundaries hold. Workflow platforms connected to ERP should be treated as part of the business systems attack surface.
Sponsor Industry-Grade Research
SAP Patches Critical ERP Risks
SAP’s July Security Patch Day brought the ERP-specific risk picture into focus.
SAP’s official July Patch Day page noted the company released 16 new security notes, one GitHub security advisory, and three updates to previously released notes. Onapsis counted the cycle as 20 new and updated SAP security notes, including four HotNews notes and six High Priority notes.
The most severe issue was CVE-2026-44747, a memory corruption vulnerability in SAP NetWeaver Application Server ABAP with a CVSS score of 9.9. SAP listed the issue as critical. Onapsis said successful exploitation by an authenticated attacker could allow unauthorized data access, data modification, or system unavailability.
SAP also patched CVE-2026-27690, an HTTP request smuggling vulnerability in SAP Approuter with a CVSS score of 9.1. Onapsis said the issue affects SAP Approuter deployments in non-Cloud Foundry environments and could lead to request-response desynchronization.
The third major issue was CVE-2026-44761, an insecure sample credentials vulnerability in SAP Commerce Cloud with a CVSS score of 9.1. SecurityWeek reported the Commerce Cloud issue involved sample configuration scripts that may have created OAuth2 clients with known credentials if customers used them in production without replacing secrets.
SAP also updated a June Security Note for CVE-2026-40128, a directory traversal vulnerability in SAP NetWeaver Application Server Java with a CVSS score of 9.0.
The spread of affected products is a reminder that SAP risk is not confined to one layer. July’s cycle touched core application server infrastructure, routing components, commerce environments, Java web components, Integration Suite dependencies, SAProuter on Windows, SAP S/4HANA authorization checks, and UI-related vulnerabilities.
SAP Patch Priority Beyond CVSS
SAP customers cannot patch by score alone.
CVSS helps identify severity, but exposure determines urgency. A critical NetWeaver issue in a tightly controlled internal environment, an Approuter flaw exposed through a web-facing architecture, and a Commerce Cloud credential issue caused by sample configuration behavior create different remediation paths.
SAP environments also carry operational constraints. Customers must test kernel updates, application dependencies, custom code, integrations, regression risk, and business process impact. Delayed patching can expose sensitive business data, but rushed patching can disrupt financial close, procurement, order management, production planning, customer commerce, or integration flows.
That puts more pressure on SAP Basis, security, infrastructure, and business-process teams to work from a shared exposure model.
The July Commerce Cloud issue is especially useful as a warning. Known credentials created from sample scripts are not just a patching issue. They are also a configuration, deployment hygiene, and secure-by-default issue. Customers need to verify whether sample assets, default accounts, temporary credentials, and development shortcuts ever made it into production.
Analysis
What this means: SAP patching has to be exposure-led. NetWeaver, Approuter, Commerce Cloud, Integration Suite, SAProuter, and S/4HANA issues do not carry the same business risk in every landscape. ERP security teams need to map vulnerabilities to internet exposure, business criticality, data sensitivity, compensating controls, and the operational window available to patch.
Get Our Free Weekly Newsletter
Patch Velocity Meets Business Risk
Microsoft, ServiceNow, and SAP are different patch stories, but the enterprise pattern is the same.
AI is increasing vulnerability discovery volume. AI-enabled platforms are becoming critical systems in their own right. ERP applications continue to expose high-value business data and operational workflows. Identity, collaboration, workflow automation, application routing, commerce, and core ERP infrastructure now sit in one interdependent risk model.
That creates a prioritization problem. Security teams may see hundreds of vulnerabilities in one cycle, while ERP teams see a smaller number of fixes that could affect the business systems they cannot afford to break. The highest-risk work is deciding what must be patched immediately, what requires compensating controls, what can wait for testing, and what needs architecture remediation beyond a vendor patch.
Enterprises should also expect patch communication to become more complex. Vendors will increasingly disclose vulnerabilities found through AI-assisted systems. Some of those flaws may have no known exploitation. Others may become exploitable faster after disclosure because attackers can use AI to analyze patches and build exploit paths.
For ERP leaders, vulnerability management requires more than a monthly review meeting. It needs current asset inventory, exposure mapping, dependency knowledge, emergency change capacity, tested rollback plans, business-process awareness, and clear ownership across cloud, infrastructure, SaaS, and application teams.
The July patch cycle shows where this is heading. The enterprises that handle vulnerability management as a shared business systems discipline will move faster and with less disruption. The ones that still treat patches as isolated IT maintenance will struggle as AI makes both discovery and exploitation move at machine speed.
Analysis
What this means: ERP vulnerability management needs a faster operating rhythm. Patch teams should know which systems face the internet, which platforms support finance or supply chain operations, which SaaS updates are already applied, and which self-hosted systems still need action. AI-era security will reward organizations that can turn vulnerability data into business-risk decisions quickly.
