COVERAGE INITIATION : Dynatrace (NYSE: DT) : Observability, and how the agentic shift is reshaping where its value sits

Why we are initiating

Observability is quietly changing jobs. For fifteen years, the category answered one question: when something breaks, why? Dashboards, alerts, root-cause. The buyer was an engineering team, and the winner was whoever explained failures fastest.

That job is not going away, but it is no longer where the money moves. The new question is: can the system be trusted to act on its own? As companies put software agents into live operations — things that don't just report a problem but take steps to fix it — the risk shifts from "did we see the problem" to "did the automatic action make things worse." Observability is being pulled from the watching seat into the controlling seat. That is a bigger, more valuable, and more dangerous seat to sit in.

This initiation covers where Dynatrace sits in that shift, what it does genuinely well, where it is exposed, and — our signature — three specific customer types it should be pursuing now that most observers are not pointing at.

---

What Dynatrace actually is

Dynatrace watches large, complicated software systems and tells the operators what's happening inside them. Its long-standing edge is automatic understanding: it maps how all the pieces connect by itself, and its reasoning engine (branded Davis) points at the true cause of a problem rather than making a human dig for it. In early 2026 it repositioned the whole platform around acting, not just watching — describing observability as the foundation an operating system for agents would need.

The important nuance, and the thing most coverage gets wrong: Dynatrace's pitch is not "we have the smartest AI." It's "we're the safe one." They argue that an agent should only be allowed to act when the underlying facts are certain — they pair rule-based certainty with the flexible AI part, and sell that combination as trustworthy automation. In a market nervous about handing control to software, "trustworthy" is a deliberate wedge. Remember that word; it defines both their strength and their trap.

---

Where the capability is genuinely strong

Automatic dependency mapping and causal analysis remain best-in-class — for a large, tangled environment, the platform figures out the connections and the true cause with less human setup than rivals. This is exactly the muscle the agentic era needs, because an agent can only safely act if the map underneath it is accurate. Dynatrace's oldest strength happens to be the new era's scarcest ingredient.

The "safe automation" position is well-chosen. By insisting that certain facts must gate any automatic action, Dynatrace aims straight at the fear that stops enterprises from deploying agents at all. For regulated buyers — banks, insurers, large retail — that framing matters more than raw cleverness.

Buyer trust is high where it's deployed. Independent enterprise reviews rate it as strongly as any competitor, and its base skews toward exactly the large, complex, high-stakes environments where the agentic-control seat is most valuable.

---

Where the capability is exposed

The cost and complexity reputation is a real drag. Across independent comparisons, two criticisms recur: Dynatrace is expensive to start, and its consumption-based pricing is hard to predict as usage grows. A wave of cheaper, simpler challengers competes specifically on "no bill shock." This doesn't threaten Dynatrace in the giant accounts — it threatens its ability to reach down into smaller ones.

The "safe/deterministic" wedge is double-edged. If the market decides "move fast" beats "move safely," Dynatrace's caution could read as slowness. Their strength is conditional on enterprises staying nervous.

Everyone is converging on the same agentic story. The nearest large rival is expanding from watching into acting too. "Observability becomes the control layer for agents" is becoming the whole category's narrative, not Dynatrace's alone. Being early to say it is worth less than being the one buyers believe.

---

Three accounts Dynatrace should be pursuing that aren't obvious

Pick 1 — Mid-sized regulated firms currently priced out, reached through a deliberately simpler tier. Dynatrace's "safe automation" story is worth most to regulated buyers — but its price and complexity lock out the mid-sized banks, insurers, and healthcare firms who feel the same fear with smaller budgets. The move isn't chasing bigger logos; it's a stripped, predictable-price entry tier for regulated mid-market buyers who want the trust wedge without the enterprise weight. Rivals own this segment on price today — but can't match the safety story.

Pick 2 — Companies deploying other people's AI agents, not building their own.** Most attention goes to firms building AI. The larger, quieter population is ordinary enterprises now running agents bought from vendors — and afraid of what those agents might do unsupervised. They don't need help building AI; they need a referee that watches a third-party agent and stops it before it causes harm. Dynatrace's watch-and-gate capability fits, and this buyer is far larger and less contested than the AI-builder crowd everyone chases.

Pick 3 — System integrators as a route, not just as customers.** The firms doing the hands-on work of putting agents into large enterprises are the ones enterprises trust to say which observability platform to standardize on. Dynatrace should pursue them as distribution — become the safety layer the integrator recommends by default in every agent rollout. One integrator relationship carries the platform into dozens of accounts it would otherwise chase one at a time.

---

What "coverage" means here

We maintain an ongoing, current view of the observability category and Dynatrace's position within it. This note is refreshed quarterly and whenever something material happens. Prior account picks are scored openly in each update — hits and misses both stay visible.

This is independent coverage. Nobody pays to be covered, and nothing here is for sale. Vendors within our coverage universe are welcome to brief us so our view stays accurate; that access informs our judgment but never purchases a position in it.

Briefings and inquiries: briefings@analystlayer.com