The Antidote to Lock-In: Can Open Standards Win the Security AI Platform Wars?

COLUMBIA, Md. – October 20, 2026 — For the past three years, the cybersecurity industry has been selling a compelling but largely unfulfilled promise: that Artificial Intelligence will cure the chronic fatigue of the Security Operations Center (SOC). Yet, for the analysts on the ground, the operational reality has been starkly different. Instead of a unified, autonomous assistant that seamlessly correlates threats, security teams have been handed a dozen fragmented AI chatbots, each trapped within its own proprietary walled garden.

Today, Tenable Holdings, Inc. is attempting to shatter those silos. The exposure management giant announced the launch of Tenable OPEN for AI, a significant expansion of its partner ecosystem designed to connect its deep exposure intelligence with third-party AI models, tools, and agents. By adopting the open Model Context Protocol (MCP) standard, Tenable is drawing a distinct line in the sand against the closed ecosystems championed by enterprise tech behemoths, betting that security leaders will prioritize best-of-breed flexibility over all-in-one consolidation.

The initiative launches with a formidable roster of partners, including BeyondTrust, Cohesity, Command Zero, Corelight, Infoblox, Recorded Future, Securonix, SentinelOne, Setyl, Snyk, Sparkle In, and Splunk. It allows organizations to use their preferred AI models to interact with datasets across these vendors simultaneously, theoretically unlocking cross-platform visibility without the need for custom software development.

Standardizing the Security Stack: Why the Model Context Protocol Matters

To understand the strategic weight of this announcement, one must look past the AI marketing gloss and focus on the architectural shift happening beneath the surface. For decades, integrating security tools meant relying on bespoke, brittle APIs that required constant maintenance.

Tenable's approach leverages Anthropic's Model Context Protocol (MCP), an open standard rapidly emerging as the "USB-C port for AI." MCP provides a universal interface layer that allows Large Language Models (LLMs) to securely connect with external data sources and tools. Through the newly announced Tenable Hexa AI MCP Server, Tenable is exposing 90 structured tools from its Exposure Data Fabric to any MCP-compatible AI client.

"Tenable is the only exposure management provider offering a truly unrestricted, vendor-agnostic ecosystem of validated AI use cases, based on customer demand," said Eric Doerr, chief product officer at Tenable. "By supporting both standardized MCP connectors and direct APIs, Tenable OPEN for AI provides security teams complete freedom to integrate their AI tools of choice and build custom, agentic workflows without artificial restrictions or forcing reliance on a single proprietary platform."

This means an organization can adopt a "bring your own AI" (BYOAI) approach. Whether a SOC uses a custom-built LLM or an off-the-shelf enterprise assistant, it can query Tenable's vulnerability data natively. These workflows are available either as documented processes through the Hexa AI MCP Server or as ready-to-use skills on the CyberAgents Exchange, a vendor-agnostic AI agent marketplace powered by Tenable.

Autonomous SOCs in Practice: Moving from Dashboards to Dialogue

The tangible impact of this interoperability is best observed in the specific use cases rolling out today. In a traditional SOC, an analyst investigating a network alert might have to pivot across four different dashboards to determine if the targeted asset is vulnerable, who owns it, and what the potential blast radius is. Tenable OPEN for AI aims to condense this into a single natural-language query.

The integration with SentinelOne, for instance, focuses on cross-platform security triage. "The ability to unite disjointed data across both endpoints and enterprise risk exposure and take coordinated, automated action transforms security teams," said Melissa K. Smith, SVP Global Strategic Partnerships & Initiatives at SentinelOne. "Our partnership with Tenable OPEN for AI enables customers to see SentinelOne's endpoint analysis and Tenable exposure intelligence side by side. Together, we're delivering a cross-platform triage engine that prioritizes risk based on context-rich, real-world insights without requiring complex custom integrations."

Similarly, the integration with Splunk addresses a universal pain point in security analytics. "Every SOC analyst knows the question that follows an alert: is this asset actually vulnerable?" noted Robert Rossetti, Partner Engineer at Splunk. "Tenable OPEN for AI answers it inside the investigation. Analysts can, for example, ask questions in plain language and get Splunk alerts back with Tenable vulnerability context attached, so they can confirm whether a critical asset is under active attack and respond faster."

Other critical integrations bridge the gap between distinct security disciplines. Snyk's integration joins developer-first application security findings with Tenable's runtime exposure data, allowing teams to quickly confirm if code-level vulnerabilities are actually deployed in production. BeyondTrust combines privileged account and entitlement data with Tenable findings to risk-rank assets based on identity context. Infoblox provides ground-truth network telemetry, while Command Zero ensures that priority exposure findings feed directly into active investigations, recording every step for reviewability.

The Hidden Risks of Cross-Vendor AI

While the operational efficiencies of a unified, agentic SOC are highly attractive, the proliferation of cross-vendor AI querying introduces complex governance and security challenges. The Model Context Protocol, while revolutionary for interoperability, reverses the familiar interaction pattern of enterprise software. Instead of clients querying servers, MCP often allows servers to query and execute actions for connected clients, creating new and less-traced attack paths.

Cybersecurity experts warn that if an MCP implementation is insufficiently secured, malicious actor-controlled inputs could reach execution environments without proper constraints, potentially leading to Arbitrary Code Execution (ACE). The Open Worldwide Application Security Project (OWASP) recently published a Top 10 list specifically for MCP, highlighting the unique attack surface it presents.

When sensitive vulnerability, identity, and network telemetry data is shared across multiple partner MCP servers and large language models, data sanitization becomes paramount. Organizations must ensure that exposure data is not inadvertently used to train public models or exposed through prompt injection attacks.

Tenable appears acutely aware of these risks. The company's architecture requires that the Hexa AI MCP Server operate against the same platform capabilities a user would access directly, meaning all actions respect existing access permissions and are fully auditable. Furthermore, agentic workflows that require data alteration demand explicit human approval. The CyberAgents Exchange also features an "AI Inspector" designed to check for classes of security issues across the model and infrastructure layers, including excessive agency and data exfiltration.

The Battle Against the Walled Garden

The launch of Tenable OPEN for AI is more than a product update; it is a strategic maneuver in the broader war for the future of enterprise security architecture. The industry is currently divided into two ideological camps.

On one side are the consolidated platforms—titans like Microsoft with its Security Copilot and Palo Alto Networks with Cortex. These vendors offer deeply integrated, AI-powered insights, provided you are willing to commit heavily to their proprietary ecosystems. They argue that tight integration natively reduces complexity and improves security outcomes.

On the other side is the open ecosystem approach championed by Tenable and its launch partners. This camp argues that no single vendor can see everything, and that forcing enterprises to abandon their existing, carefully curated security stacks in favor of a single provider is both unrealistic and dangerous.

By leaning heavily into the Model Context Protocol, Tenable is betting that Chief Information Security Officers (CISOs) will ultimately demand the freedom to connect their preferred AI tools with best-of-breed security telemetry. If the open standard approach succeeds in delivering the promised reductions in mean time to respond (MTTR) without compromising data governance, it could force the industry's walled gardens to open their gates, fundamentally reshaping how enterprise security teams operate in the AI era.