Enterprise Adoption is the Biggest Bottleneck for AI Agents | Randy Bias, Mirantis

What are some of the biggest challenges or bottlenecks the industry will face when we try to adapt to these changes? >> I mean the the biggest one I see right now is just enterprise adoption, right? It's the typical problem with the enterprise adoption curve. You've got new technology. It's got to be vetted and validated. It's got to be secured. Um, we just went through a whole process with um, answering a cyber security questionnaire from a major financial services company about an MCP server we're providing them. And you know, it's what you would expect, but also I was very impressed. Um, the cyber security team in question really understood MCP. They understood the issues. They understood AI. It was very, very good questions. But you can tell that the cyber security team is going, "Okay, this is a whole new attack surface. How do we take care of it? How do we make sure that, you know, we still are compliant with all of our regulatory requirements and so on?" And so there's kind of all that just general resistance, right? Enterprises have a hard time adopting brand new technology just like they did with cloud. So I think that's going to be one uh big point pain point. And then the other is um you know there there just seems to be a I don't want to say a lack of education but almost a finding your own way as a as a as a group about how to apply these technologies to be successful. So for example in the cloud native days before we even had the term cloud native people would try to put all kinds of workloads on AWS that didn't belong there and then there would be problems. So, I think it's similar here in that we have people going, "Okay, um, an AI agent can just come in and automate everything for me." And then when they start to get in the trenches and and and figure out what that means, they realize AI agents aren't, you know, the best for every solution. They're they're appropriate for what they're good at, which is fuzzy logic, right? Um, and you still have to have, you know, regular old normal software in place. Um, and you know, you have to figure out how to put these things together in a way where you're going to create business value. And so it's not automation for automation's sake using AI agents. It's going in and looking at the business doing a triage of your own portfolio and saying, "Hey, you know, I care about, you know, this is a place where we can use agents for direct business impact. This is another place where we can do and actually thinking through kind of the business objectives and almost looking at it through the lens of being a CEO for a new startup for example and really sort of asking yourself you know what is the business impact if we apply AI agents to this specific use case and solve this specific problem.

Enterprises aren't moving slowly on AI agents because they don't see the value — they're moving carefully because the stakes are real. Randy Bias, VP of Technology and Strategy at Mirantis, breaks down the two biggest blockers holding organizations back: security validation and the temptation to automate everything without a clear business case. Drawing on a real-world example of a financial services firm rigorously vetting an MCP server, Bias explains why the enterprise adoption curve for AI is playing out just like cloud did — and why that's not necessarily a bad thing. The key insight: AI agents aren't a universal fix. They're best suited for problems involving fuzzy logic, and success comes from identifying the right use cases with direct business impact. Read the full story at www.tfir.io #AIAgents #EnterpriseAI #Mirantis #MCP #CloudNative #AIAdoption #AIInfrastructure #AIStrategy #PlatformEngineering #AIAutomation