The Graphic Future Of IT Management

[Graphic created by prompting an LLM to create a graph model in Neo4j.]

As we approach the 2025 ServiceNow Knowledge and Atlassian Team conferences, IT management is entering a new era. The rapid adoption of AI-driven automation and the increasing use of graph-based models signal a fundamental shift in how organizations manage IT portfolios. IT management platforms (considered broadly) are evolving from forms and workflow-driven systems into systems based on intelligent, interconnected knowledge graphs that provide a real-time, holistic view of enterprise IT.

In 2024, we saw this shift take hold as vendors such as ServiceNow and Atlassian matured and promoted their graph-based approaches. ServiceNow continues to expand its Configuration Management Database (CMDB) with graph-based models to represent IT assets and dependencies more dynamically. Atlassian, on the other hand, has taken a system-of-work graph approach to model how teams collaborate and deliver value. These advancements mark a paradigm shift in IT management: the emergence of AI-powered graph-based IT operating models.

What is in the graph? All the usual things: servers, clusters, containers, applications, software, technology products, service offerings, cloud resources, endpoints, APIs … and projects, products, epics, tickets, stories, requirements, work orders, source code, packages, pipelines … and events, alerts, incidents, metrics, logs, traces, policies … everything. It may be centralized but at scale is more likely to be federated. Essentially, it is a massive digital twin of the IT organization.*

Why is this possible now? AI and generative AI (genAI) are overcoming the discovery and quality issues that have bedeviled IT management data for years, and the graph database is a superior platform for data integration. IT leaders have wanted this kind of a view since the days of the mainframe. We now have the technical infrastructure to create it and keep it current.

The fundamental question that will shape IT management in the coming years is: Who owns the graph? As organizations realize the power of interconnected IT knowledge, controlling and governing these knowledge graphs will be central to enterprise IT strategy. We are on the cusp of a struggle over the ownership, governance, and monetization of IT knowledge graphs.

Graphs Are Reshaping IT Management

Graph databases have long been used in adjacent domains such as fraud detection, social networks, and recommendation engines. Now, vendors are leveraging graphs to create more intelligent, dynamic representations of IT landscapes. The reasons for this shift are clear:

1. AI requires structured knowledge. GenAI and large language models (LLMs) require structured and contextualized data. Graphs provide a foundational knowledge model that enhances AI-driven automation, reasoning, and prediction. If unstructured data and the LLMs and vector databases that make sense of it are like flesh, graphs are the skeleton, the bones that give it structure. You need both.
2. Complexity requires relationships, not lists. IT service management (ITSM) tools originally were based on relational database technologies that struggled with the dependency-centric nature of IT management data. (Ever tried to write a recursive SQL query?) Graphs and their associated query languages are much more efficient approaches to modeling and using such information.
3. IT domains are converging. ITSM, DevOps, FinOps, SecOps, and AIOps are all converging, requiring a unified model of IT management. A graph-based control plane can interconnect these domains into a coherent system.

The Battle Over “Who Owns The Graph?”

The strategic importance of IT knowledge graphs raises a critical governance question: Who controls the enterprise’s representation of IT knowledge? There are multiple interested stakeholders:

• Enterprise architecture (EA), strategic portfolio, and CMDB owners. Understanding dependencies has always been a core objective of CMDBs, from their earliest days: If I change X, what is affected? EA teams need similar data for strategic purposes: Product A is approaching obsolescence; what is dependent on it? The technology is finally supporting these dreams, and portfolio managers need to see how it all comes together in terms of the work, the artifacts, and the costs.
• ITSM vendors. ServiceNow and Atlassian are embedding graph capabilities into their platforms, positioning themselves as the central source of truth for IT knowledge.
• AIOps vendors. Dynatrace and its competitors build dependency graphs from the operational data they manage, including OpenTelemetry traces and other dependencies. Already, customers are integrating such dependency data bidirectionally with CMDBs.
• Cloud providers. AWS, Azure, and Google Cloud maintain extensive metadata about infrastructure, services, and security configurations. They have a vested interest in controlling enterprise IT graphs; certainly, they are the origin of much of the base data for the graph.
• Security and risk management teams. As security increasingly depends on understanding complex attack surfaces, security and risk teams will demand control over IT graphs and may choose to build their own.
• FinOps, value stream management, and other IT functional areas. These teams will need direct access to IT knowledge graphs to ensure that their models remain grounded and relevant and that they again may choose to build their own.

This governance question will define enterprise IT operating models in the coming years. Organizations that fail to take a proactive stance on graph ownership risk ceding control to external vendors, winding up with the technical debt of redundant, sprawling graphs, and/or losing strategic visibility over their IT landscapes.

AI + Graph: A New IT Operating Model

The fusion of AI and graph databases is not just a technical shift; it is reshaping IT operating models. The next generation of IT management will center around real-time, interconnected knowledge graphs that allow AI-driven automation to replace traditional manual workflows. Key implications include:

• Automated IT decision-making. AI agents can detect issues and optimize performance.
• Proactive risk and incident management. Graph-based relationships enable AI to predict security issues or operational failures and recommend remediations before issues escalate.
• Enhanced developer productivity. Engineering teams will navigate IT landscapes more easily, improving DevOps velocity and reducing cognitive load.
• Dynamic IT governance. Policies can be linked directly into the IT graph information, leveraging a single source of truth and increasing assurance.

The Road Ahead: Preparing For A Graph-Based IT Future

The transition to AI + graph-driven IT management is inevitable, but organizations must take deliberate steps to prepare:

1. Define graph ownership and governance. Organizations should establish governance frameworks that clarify ownership, access controls, and stewardship models for IT graphs. They should work closely with their IT management and cloud providers to negotiate access and control over the graph data.
2. Integrate graphs across IT domains. ITSM, AIOps, DevOps, FinOps, and SecOps teams should collaborate to ensure that their respective graphs interoperate rather than become isolated silos.
3. Invest in graph + AI. GenAI is better with graphs, research shows. AI-powered tools can help populate, curate, and maintain IT knowledge graphs to prevent data decay and inconsistency. Prompts should be dynamically constructed on graph data, and agents should be continually scanning the environment, updating, and maintaining the graph for completeness and quality.
4. Invest in graph query skills. IT teams must develop expertise in graph query languages such as GQL or Neo4j’s Cypher to effectively leverage graph-based IT insights.
5. Manage graph data for quality and usefulness. IT resources and dependencies, especially in cloud-native environments, are highly dynamic. The graph, for all its utility, is an offline cache, a representation, a twin. You must 1) carefully curate what belongs in the graph (too low a level of detail will be unusable and unmaintainable) and 2) put in clear maintenance, governance, and quality protocols, as automated as possible using AI and genAI agents. Vendors such as ServiceNow are leading the way on quality assurance for this critical data.

Conclusion: IT Management Is Leading The Way For AI + Graph In Business

IT management is pioneering the adoption of AI + graph approaches, but this transformation will not be confined to IT alone. As organizations recognize the power of knowledge graphs, similar models are spreading into broader enterprise management disciplines such as HR, finance, supply chain, and risk management.

The struggle to implement, optimize, and govern the graph will shape IT strategy for years to come. Organizations that proactively define their IT graph strategy will be well positioned to harness AI-driven automation, gain real-time insights, and maintain strategic control over their IT landscapes. Conversely, those that neglect graph governance may find themselves outstripped by competitors finally getting a handle on effective IT management.

In the upcoming months, expect to hear more about AI + graph-driven IT management. The future is here, and IT leaders must decide whether to embrace it strategically or risk being left behind.

(Hat tip to my friend and world-class AI guru Dan McCreary for an excellent lunchtime conversation on these and related matters, as well as credit to Carlos Casanova for his review and feedback.)

*To be clear, the graph is metadata; it is not the actual things — fingers pointing at the moons, in the old Buddhist saying. It may, however, include embeddings (fingerprints) of the things or at least pointers to those …