IBM Links Intelligent Operations to Agent Safety and Business Value

IBM Corp. is linking its intelligent operations strategy directly to agent safety and measurable business value, according to a report from SiliconANGLE. The company argues that as artificial intelligence agents assume more enterprise workloads, organizations must establish operating models that connect automated actions to customer outcomes, infrastructure dependencies and financial accountability.

Tyler Lynch, outbound chief technology officer of IBM Automation Portfolio, stated that teams must move beyond isolated projects to consider how their work fits across the broader business. He emphasized that developers should step back to think about capabilities and outcomes for clients, noting that this approach requires multidisciplinary pillars. Lynch made these comments during an interview series on theCUBE, SiliconANGLE Media’s livestreaming studio, ahead of IBM TechXchange 2026.

Context on Agent Oversight

Lynch explained that no single approach fits every working environment when deploying AI agents. He noted that some agents operate like applications, while others are autonomous or delegated. Additionally, local co-work agents, such as coding assistants, represent different modalities that require distinct handling for authorization, logging and telemetry.

The IBM executive stated that the industry has not yet nailed down a universal standard for these varied agent types. This lack of standardization highlights the need for an operating model that can manage the diverse ways people deploy and use these systems, ensuring that oversight accounts for specific deployment contexts rather than relying on a one-size-fits-all solution.

Financial and Infrastructure Implications

IBM supports Nvidia Corp.’s Open Agent Safety Platform, contributing enterprise capabilities that span agent identity, credentials and hybrid infrastructure, according to SiliconANGLE. Lynch discussed extending agent safety into financial oversight during conversations with Apptio Inc., an IBM company. This integration aims to provide greater visibility into the costs and usage patterns associated with AI agent operations.

“We can now start to emit telemetry directly out of these agents for token usage in and out,” Lynch said. This capability allows organizations to track resource consumption more precisely, linking operational signals to financial metrics. By bringing automation, data and operational signals into a broader framework, IBM seeks to help enterprises answer the fundamental business question of what AI should help them accomplish.

What should you do?

Organizations deploying AI agents should review their current oversight mechanisms to ensure they cover various agent modalities, including autonomous, delegated and local co-work systems. Security and operations teams should evaluate whether their authorization, logging and telemetry practices can handle the distinct requirements of different agent types.

What to do next

• Audit existing AI agent deployments to categorize them by modality, such as autonomous, delegated or local co-work agents, to identify gaps in oversight.
• Implement telemetry collection for token usage and other operational signals to link agent actions to financial accountability and infrastructure dependencies.
• Review identity and credential management processes for AI agents to ensure they align with enterprise security standards and hybrid infrastructure requirements.
• Engage with industry initiatives like the Open Agent Safety Platform to stay informed on emerging standards for agent safety and control.

What to do and how to stay safe: IBM

• Sudden role changes in output signal a successful override of system instructions.
• Hidden metadata in images or documents often carries invisible injection payloads.
• Immediate isolation of the affected session prevents lateral movement within the AI workflow.

Prompt injections exploit the model’s ability to follow instructions by hiding malicious commands in input data. Isolate affected sessions immediately and implement external guardrails to filter inputs before they reach the model.

Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

What is IBM's approach to intelligent operations?

IBM ties intelligent operations to agent safety and business value by connecting AI agent actions to customer outcomes, infrastructure dependencies and financial accountability.

Who is Tyler Lynch and what is his role?

Tyler Lynch is the outbound chief technology officer of IBM Automation Portfolio, discussing the need for multidisciplinary pillars in AI deployment.

What platform does IBM support for agent safety?

IBM supports Nvidia Corp.’s Open Agent Safety Platform, contributing enterprise capabilities spanning agent identity, credentials and hybrid infrastructure.