Research Finder
Find by Keyword
Is the Dream of Self-Healing Software Finally a Reality or Just Another Costly Mirage?
Exploring Dynatrace Intelligence at Perform 2026 and whether fusing causal and agentic AI can truly automate the modern cloud-native muddle.
01/30/2026
Key Highlights
- Dynatrace Intelligence combines deterministic causal AI with agentic reasoning to move beyond simple monitoring toward autonomous system remediation.
- The updated Smartscape dependency graph and Grail data lakehouse provide the foundational context required to prevent agentic AI from hallucinating in production.
- New Intelligence Agents are designed to orchestrate complex workflows across third-party ecosystems like ServiceNow and AWS to facilitate closed-loop operations.
- Advanced Real User Monitoring now integrates frontend telemetry with backend causal context to identify how code changes directly impact customer behavior.
The News
Dynatrace used its Perform 2026 conference to announce the launch of Dynatrace Intelligence, a new system designed to coordinate autonomous actions across cloud environments. This update marks a transition from traditional AIOps to a framework where specialized AI agents can reason about and fix infrastructure issues without human intervention. The company also introduced expanded multi-cloud integrations and next-generation real user monitoring to unify observability and security data. Find out more by clicking here and here to read the press releases.
Analyst Take
We find ourselves at a rather interesting juncture where the sheer volume of telemetry has become a bit of a burden for most IT departments. The industry has spent years collecting every log, metric, and trace; however, the actual utility of this data often remains locked behind manual queries and human intuition. At Perform 2026, we see Dynatrace attempting to solve this by moving the goalposts from simple observability to what they call agentic operations. It is a bold move, particularly as it suggests that the role of the human operator is shifting from a first responder to a high-level governor.
What Was Announced
The centerpiece of the Perform event was Dynatrace Intelligence, which is architected to fuse deterministic AI with agentic AI. This is a subtle but important distinction in the current market. While many vendors are pushing Large Language Models that predict the next word in a sentence, Dynatrace is using its Davis AI engine to provide causal context. This means the system doesn’t just guess what is wrong; it uses the Smartscape topology to map out exact dependencies in real-time. The new Intelligence Agents are designed to work together across different domains; for example, one agent might identify a security vulnerability in a Kubernetes cluster while another orchestrates the remediation via a GitHub pull request or a ServiceNow ticket.
The technical specifications of the Grail data lakehouse have also been refined. It is now architected to handle even higher cardinalities of data without the need for indexing, which we see as a direct response to the rising costs of cloud logging. Furthermore, the Dynatrace MCP Server aims to deliver a standardized way for AI assistants to interact with live production data. This uses the Model Context Protocol to feed Bedrock agents or other third-party LLMs with actual facts from the production environment, which should, in theory, reduce the risk of AI-generated nonsense. The updated Real User Monitoring is also worth a mention; it is designed to tie frontend user sessions directly to backend traces and business events, allowing teams to see the financial impact of a slow page load in real-time.
We see this as an attempt to create a "digital twin" of the entire enterprise stack. The complexity of modern microservices is, frankly, a bit of a mess. By using Smartscape as a real-time dependency map, the platform aims to provide the necessary guardrails for AI agents to operate safely. Without this causal grounding, giving an AI agent the keys to your production environment would be a recipe for disaster. We have seen enough "automated" systems go rogue to be skeptical, but the Dynatrace approach of anchoring agents in environment-specific facts is a sensible way to build trust.
The market for observability is currently undergoing a massive consolidation. With Cisco’s acquisition of Splunk and Datadog’s rapid expansion into security and cloud service management, and Snoflake entering the fray with its recent acquisition of Observe, the competition is fierce. Dynatrace is positioning itself not just as a tool, but as an operating system for the cloud. This platformization strategy is designed to reduce tool sprawl, which remains a top priority for Chief Information Officers who are tired of paying for fifteen different monitoring products that don't talk to each other.
There is, of course, the question of cost. Grail is a powerful piece of engineering, but storing every scrap of data in a high-performance lakehouse can get expensive quickly. We will be watching to see how Dynatrace balances this need for total visibility with the economic realities of their customers' budgets. It is all well and good to have an autonomous system, but if it costs more than the engineers it replaces, the math simply doesn't add up. Nevertheless, the move toward agentic AI that can actually "do" things rather than just "show" things is the right direction for the industry.
Looking Ahead
Based on what we are observing, the industry is entering a post-telemetry phase where the collection of data is no longer the differentiator. The value has shifted entirely to the layer of reasoning that sits atop that data. Dynatrace Intelligence represents a sophisticated attempt to bridge the gap between human intent and machine execution. When you look at the market as a whole, the announcement signals a move away from "dashboard culture" toward "outcome-based automation."
The key trend that we are going to be looking out for is the interplay between general-purpose LLMs and domain-specific causal AI. While Datadog has been aggressive with its Bits AI assistant to improve the user experience, Dynatrace is doubling down on the "deterministic" nature of its insights. This is an epistemological battle for the soul of AIOps. Going forward, we are going to be closely monitoring how the company performs on its promise of "safe" autonomy. The adoption of the Model Context Protocol is a savvy move that acknowledges Dynatrace cannot be the only AI in the enterprise; it must instead become the source of truth for all other agents. HyperFRAME will be tracking how the company does in maintaining its lead in the high-end enterprise segment as Cisco-Splunk begins to harmonize its combined portfolio. My perspective is that the success of this agentic era depends less on the elegance of the algorithms and more on the integrity of the underlying data graph. If the map is wrong, the agent will inevitably walk off a cliff.
Steven Dickens | CEO HyperFRAME Research
Regarded as a luminary at the intersection of technology and business transformation, Steven Dickens is the CEO and Principal Analyst at HyperFRAME Research.
Ranked consistently among the Top 10 Analysts by AR Insights and a contributor to Forbes, Steven's expert perspectives are sought after by tier one media outlets such as The Wall Street Journal and CNBC, and he is a regular on TV networks including the Schwab Network and Bloomberg.