It comes up in many conversations I have with network engineers and architects these days. They want AI-enhanced management tools.
“I would like to have an agentic interface for data gathering,” a network engineer at a large financial services company recently told me. “We could just type a prompt about what is happening on the network, and the agent can pull that information for us.”
Providers of DNS, DHCP, and IP address management (DDI) solutions are not exempt from this demand. EMA recently published “DDI Directions 2026: Preparing Core Network Services for an Agentic, Multi-Cloud World,” a research report based on a survey of 300 IT pros who are responsible for their company’s DDI strategy. That research found that 43% believe that AI features are essential and are a top consideration when they are evaluating DDI vendors.
Another 53% say AI features from their DDI vendors are useful, but they don’t significantly influence vendor selection. In other words, nearly every IT organization would use AI features if their DDI vendors offered them.
This demand is outstripping availability. AI features are not widely available from DDI vendors today. Some vendors have introduced AI chatbots to their products, and others have added agentic capabilities that can analyze problems and suggest fixes. However, most DDI vendors are in the early stages of developing AI capabilities.
Chatbots are Old News; Bring on Agentic Ops
Expectations for this AI-driven DDI capabilities are also high. Chatbots aren’t good enough for most engineers. Only 10% are looking for an informational chatbot that delivers data or product information based on a query. Another quarter of engineers want an active chatbot that makes automated changes to the DDI platform based on a user’s prompt.
Instead, nearly 66% want agentic AI operations features that can actively predict or detect problems and respond to them. Nearly 40% say they want an agentic AI feature that recommends fixes to problems and lets human operators actually do that work. The rest (26%) want partially or fully autonomous AI that remediates some or all problems without human involvement.
Agentic AI is beyond what is generally available from DDI vendors today. I expect we’ll see such features hit the market over the next 6 to 12 months. Some DDI architects recently told me that they have received roadmap briefings from their vendors, and I have received a few such briefings, too.
Evaluate Agentic DDI Solutions Carefully
AI is prone to mistakes. As DDI vendors introduce AI features, engineering teams must be prepared to evaluate them. EMA published research earlier this year on AI-driven network management in general. In that report, 47% of early adopters of AI-driven network management solutions said that their AI tools provide false or mistaken insights and recommendations somewhat to very often.
“I’m torn on AI,” a DDI engineer with a multinational bank said. “I’m concerned about the accuracy and how we can vet that. If AI starts to hallucinate on source of truth data, it’s going to cause a lot of grief. There must be a way to validate it.”
Network engineers must be ready to evaluate AI tools introduced by their DDI vendors. And they must be vigilant against mistakes.
