ATLANTA, October 27, 2026 / PRNewswire / -- At TechXchange 2026, IBM ( NYSE : IBM ) announced the launch of a new feature to help developers, IT teams, and security teams use AI proxies within business systems, while maintaining clear definitions of permissions and responsibilities. IBM integrates proxy discovery, verified identity, access control, and performance assessment to help enterprises understand the ownership of each proxy, control the content they can access, and track their behavior within existing applications and infrastructure.

According to the data from Gartner, by 2028, companies comprising the S&P 500 index are expected to use an average of over 150,000 AI proxies. However, only 13% of these organizations believe that they have established appropriate AI proxy governance mechanisms [1]. Persistent credentials issues, unclear ownership, and fragmented control measures make it more difficult for companies to manage these proxies when they interact with each other, utilize tools, and operate across teams, platforms, and cloud environments.

IBM, Senior Vice President of Software, stated: "The winners in the AI industry are changing the way businesses operate, not just adding more AI to their operations. To enable agents to perform actual work within the various systems that businesses rely on, a AI operational model is needed – one that connects agents with the enterprise's data, applications, and infrastructure, and operates them with the same level of discipline as the most critical systems. Once trust and control are embedded into this model, agents cease to be merely technology; they become a way of doing business."

IBM Today's announcement expands its product portfolio, aiming to provide an AI operational model that consists of four core systems working together: agents, data, automation, and hybrid infrastructure. The starting point is the trust and control required for each agent – verified identities, access rights limited to what is necessary, as well as a platform that allows for unified discovery, protection, and evaluation of agents at any location of construction.

Expand the token issuance from Hash to rp and Vault to AI agents.

Hash Token Issuance rp Vault is used to manage the keys and credentials required for accessing business systems of applications and services. It provides control by centrally managing keys, automatically rotating supported credentials, and generating credentials that can be dynamically created, expired, or revoked. These capabilities help to reduce the time during which credentials can still be used after they are exposed, and limit access within the scope of permissions configured for each role.

Hash Token Issuance rp Vault has already provided non-human identity protection for institutions such as ABN AMRO, Adobe, athenahealth, BBVA, BlackRock, Canva, Duke Energy, Expedia, the London Stock Exchange, Starbucks, the Swiss Federal Railways, and Target. The transaction amounts processed by these systems amount to billions of dollars [2]. Hash Token Issuance rp will now extend these capabilities to AI agents, issuing them protected, short-term valid credentials, and restricting each agent to within their required scope.

Establishing Proxy Identity and Attribution of Responsibility

AI Agents are increasingly becoming a growing category of non-human identities, on par with applications and services. As they begin to perform operations within business systems, companies need to keep clear records of the owners of each agent and the permissions granted to them. Hash Token issuance rp combines governed agent identities with credential control to help companies assign tasks within a clear framework of permissions and responsibilities.

When used in conjunction with the Hash token issuance and rp Vault, Agent Identity will establish authoritative records for each agent, such as their owner, lifecycle status, and granted permissions. This solution aims to expand existing identity and security investments, and share relevant identities and authorization contexts with the credential control of Vault. Ultimately, it achieves minimal permission access and supports real-time revocation when agents attempt to perform unauthorized operations.

Control Proxy Credentials and Access

Identifying proxies is just the first step. Enterprises also need to control the credentials used by these proxies to access databases, applications, and cloud services, as well as the validity period of those credentials. As proxies invoke tools and operate across business systems, their credentials become a critical security boundary. If proxies possess long-term valid credentials with excessive permissions, they can pose risks to systems and data if compromised or deviate from their designated tasks. Enterprises must clarify the responsibilities of each proxy, limit the access rights granted to them, and revoke these permissions when circumstances change.

These controls are not limited to initial authentication. If a proxy attempts to obtain information beyond its authorized scope or perform unauthorized operations, the system can instantly deny access and revoke the session to prevent unauthorized activities. Short-term credentials and minimal permission controls for Vault can further restrict the content that proxies can access and the duration of those permissions.

Other new capabilities include:

• Vault 2.2 provides exclusive protected accounts for proxies, automatically renews their credentials, and incorporates post-quantum encryption to address the threats of the quantum era.
• Secure Hub (Public Preview) provides enterprises with centralized visualization across Vault environments, facilitating the understanding of operational status, usage, and security resources.
• athenahealth: Use the PKI engine of Vault for automated certificate management.
• Albert Einstein College of Medicine: Use IBM Verify and Vault for AI proxy identity, immediate access, dynamic keys, and auditability.
• Optum: Use Vault Radar to detect exposed keys, and use Boundary to implement identity-based instant access.
• Seven - Eleven Japan: Use Vault Enterprise to centralize and encrypt keys, and leverage Vault Radar to detect hardcoded keys.
• Yum ! Brands: Building an identity security platform with the help of IBM Verify, and looking forward to machine identity and key management.

Albert Einstein College of Medicine Deputy Dean of Information Technology Shailesh M . Shenoy states: " AI brings extraordinary opportunities for accelerating biomedical research and discovery, but to realize its potential, a foundation of security, trust, and responsible governance is required. Based on my experience, IBM Verify and Hash token issuance rp Vault and other technologies can play an important role in helping organizations establish the necessary identity, access, and security controls, making innovation more confident."

Cross-platform management proxy

Identity can tell a company what each agent is, but companies also need to see every agent they possess, regardless of where it is built. IBM watsonx Orchestrate can operate across the AI ecosystem, rather than being limited to a single supplier platform, thus providing a unified place for companies to discover, protect, and evaluate agents. Its capabilities include:

• Discovery (Technical Preview): Enterprises are unable to manage proxies that they are not aware of. Orchestrate can connect to platforms such as Amazon Bedrock, Microsoft Foundry, and Google Gemini, allowing for the detection of proxies built within third-party environments. This enables teams to see and manage these proxies without the need for manual addition. This capability is also extended to models and MCP servers.
• Isolation (Technical Preview): A rogue agent may interfere with other agents and spread errors throughout the business. Orchestrate Now, agents are run in isolation sandboxes to prevent rogue agents from interfering with others. The team can also use “kill switch” to shut down any agent.
• Evaluation: Untested and untracked proxies may deviate from the intended goals and increase costs. Teams using Orchestrate can now test the quality, accuracy, and compliance of proxies, models, and MCP servers before deployment, and continuously monitor them in a production environment. Built-in cost tracking displays the usage of models for each proxy and token, helping teams identify cost sources and make necessary adjustments.

Orchestrate can also be used in conjunction with the credential control of Vault and Agent Identity to clarify the responsibilities of each agent, govern their permissions, and implement least privilege access using credentials. By combining these controls, enterprises can authorize and audit agent behavior, and quickly revoke trust and credentials when risks change.

Connect the proxy to the business.

Once agents are authorized to use the company's API and tools, they also need to know what existing systems there are, how the systems are connected, which content can be safely reused, and what the implications of their operations may be. IBM Integration Intelligence Hub is designed to provide this information through its Unified Context Engine, enabling agents to utilize existing systems to fix stuck messages or deploy API. Skills, approval policies, and audit controls will govern these operations. Enterprises do not need to rebuild systems for AI; instead, they can allow agents to work within the systems they already rely on.

Leverage NVIDIA to expand the scale of proxies

Agent Identity and Hash tokens can be integrated with Vault, which in turn is based on NVIDIA OpenShell Architecture and rp. When combined, they provide a protected operating environment for proxies, verified identities for each proxy, as well as short-term credentials for each one, enabling every operation to be tracked, restricted, and audited.

NVIDIA, Vice President of Enterprise Software, stated: "Enterprises need to have executable control over the way proxies operate and the content they can access. Our collaboration with IBM combines the policy governance runtime of NVIDIA OpenShell with Agent Identity and Hash token issuance rp Vault, to help enterprises extend AI proxies with secure execution, verified identities, and strictly limited business system access."

Coping with Emerging Security Threats

When new vulnerabilities are discovered, enterprises must quickly identify the affected systems, determine the necessary changes, and deploy fixes without disrupting critical operations. In complex environments, this process often involves multiple teams, tools, and procedures.

IBM Concert integrates observability, optimization, event management, and security capabilities at the application, infrastructure, and network levels. With the insights and availability assurance driven by AI, Concert helps teams identify high-risk vulnerabilities, assess operational impacts, and coordinate responses. These capabilities also assist enterprises in addressing performance issues, managing costs, and maintaining resilience in hybrid environments.

Hash Token Issuance rp Terraform Provides the infrastructure context and governed workflow for change implementation. New features are now officially available or in public preview, including Terraform policy, Terraform Developer Portal, Terraform remote MCP server, enabling developers and AI agents to access approved configurations and implement infrastructure changes through established controls. Terraform driven by Infragraph also helps teams understand the relationships and dependencies between infrastructures, thereby allowing for more informed decision-making.

When combined, these capabilities enable enterprises to progress from identifying vulnerabilities to implementing and verifying fixes, while also optimizing the visibility and governance required for business operations.

Intelligence and Change Orchestration Layer for IBM Z

IBM Z Atlas can map the entire technology stack of the customer, including infrastructure, middleware, and applications. This information is then utilized to handle manual tasks such as patch management, regulatory responses, and sharing system knowledge between system programmers and application developers. Atlas by mapping how changes affect the technology stack and testing and verifying them, has addressed the primary cause of unplanned downtime – unproven-safe changes. This design is currently in the technical preview phase, which allows for accelerating the customer's modernization process while maintaining expert control. It also enables operations or development personnel to share the same visual view of the Z system, thereby reducing risks.

Availability

Hash Token Issuance rp Vault 2.2, Agent Identity, watsonx Orchestrate Evaluation Capability, IBM Concert as well as Hash Token Issuance rp Terraform policy are now officially available. Secure Hub, Terraform Developer Portal, Terraform remote MCP server and IBM Z Atlas are in the public preview phase. IBM Integration Intelligence Hub as well as watsonx Orchestrate Discovery and Isolation Capability are in the technical preview phase.

Statements regarding the future direction and intentions of IBM may be changed or withdrawn without prior notice, and they represent only the goals and objectives.

More information

IBM TechXchange is the annual conference for developers and technology professionals organized by IBM, bringing together approximately 6,500 attendees from 85 countries. This year, the event incorporated the Hash token issuance nf, uniting the Hash token issuance rp community with professionals from IBM, Red Hat, and Confluent. Attendees had access to over 1,000 sessions and activities, including 180 hands-on workshops, 260 customer-led sessions, and more than 50 certification exams. Speakers included American Airlines, Bradesco, Citi, CVS Health, Deutsche Bank, DoorDash, Hertz, and US Bancorp, who shared insights on how to make AI suitable for enterprises.

Live meetings and replays can be viewed here: ibm.biz / txc26-livestream.

Supporting materials

• Enhance the security of proxy-based AI through the new capabilities of Vault, Verify, Boundary, and Secure Hub.
• IBM Agent Identity is now officially available: Building trust for the era of proxies
• Vault Enterprise Promoting Identity Security, PKI and Encryption Agility
• Announcement of the public preview of Secure Hub: Large-scale governance of Vault
• IBM watsonx Orchestrate Promoting AI Proxy Management with New Capabilities
• IBM Integration Intelligence Hub provides the context necessary for AI to take action, not just access rights.
• IBM Concert Platform Promotes Cross-Domain Proxy IT Operations
• Terraform Promoting Infrastructure Lifecycle Management in the Proxy Era
• Introducing IBM Z Atlas: The intelligence behind changes in IBM Z

[1]: Gartner "Six Steps to Identifying and Managing the Surplus of AI Agents"

[2]: Hash Token Issuance, rp, Vault Customer Cases

About IBM

IBM is a leading provider of global hybrid cloud, AI, and consulting services. The company helps customers in over 175 countries and regions to leverage data insights, streamline business processes, reduce costs, and gain a competitive advantage in their respective industries. Thousands of governments and enterprises rely on IBM's hybrid cloud platform and Red Hat OpenShift in critical infrastructure sectors such as financial services, telecommunications, and healthcare to drive digital transformation quickly, efficiently, and securely. IBM offers breakthrough innovations in AI, quantum computing, industry-specific cloud solutions, and consulting, providing customers with open and flexible options. All of this is built on IBM's long-standing commitment to trust, transparency, responsibility, inclusiveness, and service. For more information, please visit www.ibm.com.

Media contacts: Sarah Benchaita, IBM, [ email protected ]; Kate Lehman, IBM, [ email protected ]