Last weekend, my wife and I kayaked through the Uwharrie Mountains in North Carolina, following a portage trail that wound past the Doershuk archaeological site. As I picked my way along a rocky outcrop, something caught my eye: a small piece of flaked rhyolite, clearly worked by human hands centuries ago. I turned it over in my palm, running my thumb along its razor-sharp edge. The craftsmanship was remarkable—whoever had shaped this piece understood exactly what they were working with and what they needed to accomplish.
Rhyolite is a volcanic glass that fractures predictably, creating edges sharper than surgical steel. But it only becomes a useful tool when someone with skill and purpose shapes it correctly. The raw material has potential, but without the right technique and understanding, it's just another rock on the ground.
Standing there in the North Carolina mountains, holding that ancient tool, I couldn't help but think about conversations I'd had the week before in Las Vegas at Knowledge 2025 about new CMDB features. Like those ancient toolmakers, our customers are always seeking better ways to shape their raw materials—configuration data—into tools that actually help them get work done.
The Raw Material: Data That Cuts Through Complexity
Configuration data, like rhyolite, has inherent properties that make it valuable. But just as volcanic glass needs to be worked by skilled hands to become sharp and useful, our CMDB data only becomes truly powerful when it's shaped by intelligent processes and thoughtful design.
The new AI Agent Workflows for CMDB in ServiceNow's Q2 release address this challenge head-on. Just as ancient toolmakers developed specific techniques for working with different materials, these AI agents bring specialized expertise to common CMDB challenges.
Every enterprise needs a CMDB that has high-quality data management and governance to stand up to regulatory requirements. A new agent workflow sets agents to task to (1) identify the most important CI classes, (2) establish and enforce lifecycle and certification, and (3) evaluate and improve ownership data.
A second Agent Workflow for "Create a Configuration Item" workflow acts like a master craftsperson guiding an apprentice. Most CIs should be created by automated sources. However, for those few—but critical—logical elements, the agents don't just dump you into a form with dozens of fields—they understand the relationships between different CI classes, know which attributes are critical for downstream automation, and help prevent the kind of sloppy data entry that creates problems months down the road. Think of it as having several expert flint-knappers looking over your shoulder, making sure you're using proper technique from the start.
The "Search the CMDB" workflow takes a different approach to the same core challenge. Instead of requiring users to become experts in ServiceNow's query language, it lets them describe what they're looking for in natural language. The AI agent interprets the intent, navigates the relationships, and returns insights that would typically require deep knowledge of your CMDB structure. It's like having a tool that adapts its shape to fit whatever job you're trying to accomplish.
Consistency in the Craft: Service Graph Connectors
Of course, even the most skilled craftsperson needs reliable raw materials. This is where Service Graph Connectors (SGCs) come in. Just as ancient toolmakers sought out the best obsidian or flint quarries, SGCs ensure you're getting consistent, high-quality data from our various discovery sources.
The Q2 enhancements to connectors like Trellix represent an ongoing commitment to this principle. Each connector doesn't just dump data into your CMDB—it normalizes it, standardizes the field mappings, and ensures that a server discovered by one tool looks the same as a server discovered by another. This consistency is what makes the difference between data that's merely collected and data that's actually useful.
The relatively recent Service Graph Connector central interface gives you a single pane of glass from which you can easily track the quality of data flowing in from connectors, spot any issues, or get help setting up a new connector.
Think about it: if ancient toolmakers couldn't rely on their raw materials to fracture predictably, they couldn't develop the sophisticated techniques that created those incredibly sharp, durable tools. Similarly, if our CMDB data isn't consistent and reliable, all the AI agents and smart workflows in the world won't help us make better decisions.
The Master's Touch: Editable Unified Maps
But here's where the parallels got really compelling: even with perfect raw materials and proven techniques, the best tools are often those that have been customized for specific purposes. A hunting point needed different characteristics than a scraping tool, even if they're made from the same piece of rhyolite.
This is exactly the thinking behind the new Editable Unified Map functionality. ServiceNow has always excelled at automated discovery and relationship mapping—that's your basic toolmaking technique. But every organization has unique operational realities that don't always match what automated discovery reveals.
Maybe you have a critical application dependency that doesn't show up in network scans. Maybe you need to represent a business relationship that exists in practice but not in your monitoring tools. The Editable Unified Map lets you bring your human expertise to bear on these edge cases, shaping the tool to fit your specific needs.
The drag-and-drop interface feels intuitive because it mirrors how we naturally think about relationships. You can visually see the dependencies, add the missing connections, and remove the ones that aren't actually relevant to your operations. It's like being able to adjust the grip on your tool or reshape the edge for a particular task.
What's particularly clever about this implementation is that it doesn't interfere with your automated discovery processes. The system maintains a clear distinction between relationships discovered through Service Mapping and those added through human insight. This means you get the best of both worlds: automated consistency where it works, and human expertise where it's needed.
Sharp Data, Sharp Skills
The more I thought about that piece of worked rhyolite, the more I appreciated what it represented. Some anonymous craftsperson had looked at a raw material, understood its properties, and applied learned techniques to create something genuinely useful. They weren't just randomly chipping away—every strike was purposeful, informed by experience, and aimed at a specific outcome.
That's exactly what these new CMDB features are designed to help you do with your configuration data. The AI agents embody years of ServiceNow's experience with CMDB best practices, helping us avoid common mistakes and apply proven techniques. The Service Graph Connectors ensure you're working with reliable, consistent raw materials. And the Editable Unified Map lets you apply your own expertise and organizational knowledge to create tools that are perfectly suited to our specific challenges.
The result is configuration data that's not just collected, but crafted—shaped by intelligent processes and human insight into something that can actually cut through the complexity of modern IT environments.
Standing in those North Carolina mountains, I was struck by the continuity of human problem-solving. Whether we're shaping volcanic glass or configuration data, the fundamental challenge remains the same: taking raw materials with potential and turning them into tools that help you accomplish your goals more effectively.
The ancient toolmaker who created that rhyolite flake would probably recognize what we're trying to do with these new CMDB features. We're still seeking the best materials, developing better techniques, and crafting tools that are sharp, reliable, and fit for purpose. The technology has changed, but the human drive to build better tools remains exactly the same.
