Intelligent operations begin with a business question: What should artificial intelligence help an enterprise accomplish? As agents take on more work, organizations need an operating model that connects their actions to customer outcomes, infrastructure dependencies and financial accountability.
IBM Corp.’s approach brings automation, data and operational signals into that broader framework. Moving beyond isolated projects requires teams to consider how their work fits across the business, according to Tyler Lynch (pictured), outbound chief technology officer of IBM Automation Portfolio.
“First off, instead of just being a developer or whatever that is, take a step back and think about capabilities and outcomes that you want to deliver for your customers and your clients,” he said. “That’s where we should all be starting. It’s going to have multidisciplinary pillars, and that’s what we should all be doing.”
Lynch spoke with theCUBE Research’s Dave Vellante and John Furrier for IBM’s AI Operating Model interview series, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed agent controls, infrastructure context and measuring AI value ahead of IBM TechXchange 2026. (* Disclosure below.)
Extending agent safety to financial risk
Building on IBM’s focus on enterprise AI orchestration and production readiness, agent oversight must account for how people deploy and use these systems. No single approach fits every working environment, Lynch explained.
“We have agents that are deployed that are operating like applications, and some are autonomous, some are not and some are delegated,” he said. “But then we have local agents, what I would consider co-work agents. This could be your coding agent. Those are very different modalities from authorization, logging and telemetry, and no one’s nailed it yet.”
IBM supports Nvidia Corp.’s Open Agent Safety Platform, contributing enterprise capabilities spanning agent identity, credentials and hybrid infrastructure. Lynch’s discussions with the team at Apptio Inc., an IBM company, have also raised questions about extending agent safety into financial oversight.
“We can now start to emit telemetry directly out of these agents for token usage in and out,” he said. “Why can’t we have policies evaluate that just like we do for [identity threat detection and response] and maybe revoke an agent’s [application programming interface] key so it can stop? Put it in the corner until a human can go look at it.”
Giving intelligent operations shared infrastructure context
Acting on operational signals also requires understanding the environment an agent might change. HashiCorp Inc.’s Infragraph is now available to eligible Terraform customers on a limited basis.
“Infragraph is a graph-based representation of your infrastructure, but we’re extending that to be more of a knowledge graph,” Lynch said. “It’ll actually be used by our FinOps tools and our observability tools, and they will also enrich that graph.”
Shared infrastructure context could help teams assess changes beyond the resources a particular configuration directly manages. Other applications or users may rely on those resources without the team making the change seeing those dependencies, Lynch explained.
“I like the more complex use cases, which is catching the errors that humans might not catch,” he said. “Dependency impact analysis is something I’m very much interested in as a scientific area. Can we do dependency impact analysis to determine if [a change] would adversely affect anybody else?”
Measuring outcomes beyond token consumption
For intelligent operations, spending alone can’t show whether AI delivers value. Enterprises need to connect that investment to the outcomes of the processes they’re changing, Lynch explained.
“For me, speed of business matters, and that’s really hard to quantify sometimes,” he said. “I talk to people [asking], ‘Can we get something to market faster?’ That’s not just software development; that is maybe an underwriting process. It’s not just, ‘Did I fulfill an order?’ It is, ‘Did I do it faster? Did I do it better?’”
The financial operating practices supporting those decisions also develop in stages, moving from retrospective cloud bills toward shorter optimization cycles and forecasting based on actual spending. Those practices can extend operational information to FinOps teams, controllers and chief financial officers. For enterprises preparing for IBM TechXchange, Lynch emphasized thinking beyond individual roles and focusing on customer outcomes.
“How do we inform them all? You can’t do that all on day one,” he said. “We land, and then we expand.”
Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of IBM’s AI Operating Model interview series:
(* Disclosure: TheCUBE is a paid media partner for IBM’s AI Operating Model interview series event. Neither IBM, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)