Forrester spots context graphs

Enterprise AI is drifting away from the model wars and toward plumbing. That’s the real shift here. On May 1, Forrester argued that Atlassian and ServiceNow are pulling ahead because they already own the work systems, data relationships, and governance layers that AI actually needs to do useful things inside big companies. The new buzzword is “context graph,” but the deeper point is simpler — AI gets much better when it understands how people, services, tickets, assets, policies, and decisions connect. (forrester.com) ### What is a context graph? It’s a structured map of an organization’s operating reality. Not just documents and chats, but the links between a service, its owner, the incidents tied to it, the policies around it, the change tickets that touched it, and the decisions people made along the way. That’s why Forrester treats this as more than search. The graph is supposed to give AI something closer to a model of the business, not a pile of disconnected text. (forrester.com) ### Why are people talking about it now? Because generic copilots hit a ceiling fast in the enterprise. They can summarize text, but they often miss the actual relationships that make work legible. Forrester’s April note made the case that “context graphs” are really a convergence of older disciplines — CMDBs, process mining, arch(forrester.com 1)(forrester.com 2) ### What did ServiceNow actually do? ServiceNow made this explicit on April 9. It announced Context Engine as part of a broader push to make every product AI-enabled, with data connectivity, workflow execution, security, and governance built in. The company says Context Engine connects relationships, policy, and decision history behind every AI-agent action. In plain English — it wants AI to know not just what exists, but what it’s allowed to do and why. (investor.servicenow.com) ### What’s Atlassian’s version? Atlassian has been building toward this with Teamwork Graph, the data layer under Rovo. Atlassian describes it as a permission-aware map of work, connecting people, projects, knowledge, incidents, and follow-up tasks. The pitch is that Rovo doesn’t have to guess what “Checkout” or “Q3 Reliability” means inside your company, because the graph already links those phrases to the right services, initiatives, docs, and teams. (community.atlassian.com) ### Why does scale matter so much? Because the moat here is not just model quality. It’s accumulated operational context. Forrester points to ServiceNow’s 85 billion workflows and 7 trillion transactions annually, and says Atlassian’s Teamwork Graph has passed 100 billion objects and connections. Those numbers are messy and not directly comparabl(community.atlassian.com)ually don’t have. (forrester.com) ### Is this just another name for a knowledge graph? Not exactly. The catch is governance and action. A traditional knowledge graph helps software understand relationships. A context graph, at least in this framing, also carries permissions, workflow state, policy, and decision history. Basically, it’s meant to support AI that can(forrester.com)ermission-aware. (investor.servicenow.com) ### Why are Atlassian and ServiceNow the names to watch? Because both already escaped their old boxes. ServiceNow is no longer just ITSM. Atlassian is no longer just dev tools and docs. Forrester’s point is that they’ve spent years becoming broader enterprise service platforms, so they already host the workflows where AI can be grounded in real business objects instead of floating above them as a chatbot layer. (forrester.com) ### What’s the bottom line? The market is starting to reward AI systems that know what the business is, who owns what, and what actions are allowed. That favors platforms with deep context, not just flashy models. If Forrester is right, the next enterprise AI winners won’t be the tools with the best demo prompts — they’ll be the ones with the best map. (forrester.com)