Enterprises that have gone the build-your-own route with agentic AI—picking their own models and standing up their own agent infrastructure on platforms like Amazon Bedrock AgentCore—have had to figure out governance themselves. ServiceNow is offering them a shortcut. The company on Wednesday connected its AI Control Tower with AgentCore, giving customers a single layer to govern agents they’ve built across AWS, regardless of which models are running underneath.
ServiceNow is also pairing its AI Specialists with AWS agents to handle workflows across security, IT operations, and telecommunications, with humans approving key decisions. Separately, the company is bringing its SDK into Kiro, AWS’s agentic IDE, so developers can build ServiceNow applications and AI agents without leaving the editor.
“Organizations aren’t experimenting with AI anymore, they’re operationalizing it,” Chris Grusz, AWS’s managing director of technology partnerships, says in a statement. “ServiceNow and AWS are delivering the architecture to make that real: unified governance, trusted infrastructure, and developer tools that take AI from idea to impact…We’re making it easier than ever to deploy, govern, and scale AI agents across the enterprise.”
ServiceNow’s AI Control Tower debuted at Knowledge 2025 as a single pane of glass for the AI agents and models running across an enterprise. Since then, the company has added new capabilities for discovering, observing, governing, securing, and measuring those deployments. For enterprises building agents on AWS, the Bedrock AgentCore integration provides a consolidated offering rather than piecemeal assembly.
According to Joe Davis, ServiceNow’s executive vice president for platform engineering and AI, enterprise customers don’t put all their eggs in one basket—they have solutions spread across multiple hyperscalers and products. But in order to really secure all of the organization’s AI assets, it’s critical to first take inventory. “You need to have visibility of all of it to make sure that you can contain exposure,” he says in an interview with The AI Economy. “In order to do that, you have to be cross-platform, neutral, and integrate better.”
It’s why ServiceNow made AI Control Tower extensible, integrating it with Amazon Bedrock APIs to identify AgentCore-created agents. Once they’ve been discovered, the AI Control Tower can begin securing and governing them.
When it comes to AI Specialists, ServiceNow is connecting them with AWS’s AI agents so they can work more closely together. It’s a recognition that both platforms have limits in what their tools can do, so to help workers complete tasks, ServiceNow and AWS have worked on a way to ensure seamless system handoffs to minimize risk and reduce delays.
“It’s like an MCP and A2A integration, where there are parts of the workflow that our AI Specialists handle, and then they will work with other agentic systems, like AgentCore, to do some of those domain-specific tasks that maybe…our AI Specialist doesn’t do,” Davis says.
How this works is when an AWS service detects an issue, ServiceNow’s AI Specialist will orchestrate the response, a human approves the fix, and the workflow closes with a full audit trail. In security, a configuration change in the configuration management database (CMDB) would trigger ServiceNow’s Vulnerability Resolution AI Specialist to call the AWS Security Agent via MCP to run an on-demand penetration test. In doing so, it would include identity exposure data from ServiceNow’s Veza and Armis device data before remediation.
For IT operations, an anomaly in CloudWatch would be routed to ServiceNow’s AIOps and Site Reliability Engineering Specialists, which works with the AWS DevOps Agent to correlate events and execute fixes.
“It’s kind of how people work together,” Davis explained. “I have some general knowledge, but then there are subject matter experts that I might want to go ask for help. It’s kind of the same thing from an AI standpoint.”
ServiceNow is also bringing its SDK into Kiro, AWS’s agentic IDE, where developers can vibe-code ServiceNow AI agents without switching tools. The company’s SDK with Build Agent skills can now be installed from the Kiro Power Marketplace with a single click. The promise is that developers will be able to “scaffold” applications, configure workflows, and create agents through natural-language prompting, without leaving their IDE.
The expanded integration lands as ServiceNow and AWS disclosed a commercial milestone: customers have now transacted more than $1 billion through the partnership on AWS Marketplace. ServiceNow described it as something more significant than “a commercial threshold,” framing it instead as a shift in the enterprise to consolidate infrastructure around platforms they trust. ServiceNow is positioning the billion dollars as market validation of the thesis that organizations prefer a layered tech stack over point solutions or building it all themselves.
Davis highlighted that ServiceNow is focused on delivering turnkey, outcome-oriented solutions—prebuilt AI that plugs directly into real enterprise workflows with permissioning, security, and governance baked in. Doing it yourself is “extremely complicated, and there are so many things that can go wrong,” he pointed out.
ServiceNow AI Control Tower with Amazon Bedrock AgentCore is available today in the AWS Marketplace. The company’s Vulnerability Resolution AI Specialist, AIOps AI Specialist, and Site Reliability Engineering AI Specialist integrations are expected to be available later this year. Finally, the ServiceNow SDK for Kiro is available now in the Kiro Power Marketplace.
Disclosure: I attended ServiceNow’s Knowledge 26 as a guest of the company, with my flights and hotel stay paid for. The AI Economy’s coverage is editorially independent from those that it covers. These words are my own.
Thanks for reading The AI Economy! This post is public so feel free to share it.