A customer clicks a button. Somewhere beyond the browser, several pieces of software begin talking to one another. If the page freezes, the customer sees one failure. The engineers inherit a much less tidy question: which conversation went wrong? Bernd Greifeneder made that question the basis of a company. More than twenty years after co-founding Dynatrace, he is still following the consequences of the click.
It is an unusually durable obsession in a business that regularly changes its vocabulary. Application performance management became observability. Cloud systems multiplied the moving parts. Artificial intelligence introduced software that could propose, and increasingly perform, its own actions. Through those changes, Greifeneder has remained close to a practical difficulty: knowing what a system is doing well enough to intervene intelligently.
His first chief executive role at Dynatrace lasted three years. His technical work there has lasted more than twenty. He was the original company’s chief executive until 2008; technology became his continuing assignment. Dynatrace’s present leadership biography identifies him as CTO and founder, and credits him as a named inventor on twenty patents. The questions have kept him occupied.
The shop could fail. Then what?
Before Dynatrace, Greifeneder worked at Segue Software. Between January 1998 and February 2005, his roles included project lead, chief software architect and CTO of Global Technologies. The work put him near a useful source of business ideas: customers who had discovered that owning software and understanding its behavior were separate accomplishments.
Load testing could expose a problem before an online service went live. Simulate enough customers and you could discover that the system would struggle. That was valuable information. It also left an awkward pause in the conversation. The test had found the limit; someone still needed to explain it. A warning gives an engineer work. An explanation gives that work a direction.
Greifeneder wanted to follow activity through the systems involved, from the browser through servers to the database. A click looked singular to the user and became plural behind the screen. His research question concerned that hidden journey. When he returned to finish his master’s degree, he built the prototype that helped persuade him to start Dynatrace.
A visible delay can begin far from the screen. Following the request helps connect the symptom to its cause.
The name arrived in the Linz company register in early 2005. The founders were Greifeneder, Sok-Kheng Taing and Hubert Gerstmayr. Their backgrounds covered technology, marketing and ecommerce, and finance. It is a pleasingly practical founding arrangement. Software requires someone to build it, someone to sell it and someone to keep an eye on what all that enthusiasm costs.
A founder learns to divide the work
The early company demanded more than an elegant diagnosis. Greifeneder was chief executive while remaining involved in building the product. Hiring, onboarding and selling also needed attention. He later singled out the period at roughly twenty to thirty people as particularly difficult. The job had expanded in several directions at once, with the same person still standing in the middle.
His lesson was to recruit ahead of the next stage, thinking about who the business would need a year or two later. There is a modest but consequential change of perspective here. A founder may be able to solve today’s problem personally. Repeating that arrangement as the company grows turns personal competence into a queue.
His chief executive tenure ended in 2008. He continued as CTO of the original dynaTrace Software GmbH until December 2014, when his tenure as CTO of the present company began. In April 2024, he became an executive vice president. Those dates describe a long technical career inside an organization whose corporate form and scale changed around him.
“You can always focus more, it’s true.”
Bernd Greifeneder
The handover also makes his advice about teams more concrete. Responsibility has to move if an organization is to keep moving. For a founder whose interest began with software architecture, concentrating on technology offered a continuing place to apply that expertise. The company could develop its leadership without requiring its original engineer to lose interest in the product.
Keeping the engineering in Linz
Dynatrace’s journey acquired a Wall Street milestone in 2019, when it listed on the New York Stock Exchange as DT. Its origin remained firmly Austrian. The company began in Linz, and its engineering headquarters continued there. Going global did not make the place where the technical work happened an incidental detail.
Greifeneder’s education also ties him to the city. He holds bachelor’s and master’s degrees in computer science from Johannes Kepler University Linz. Earlier studies included mechanical and industrial engineering, followed by mechatronics at university. The route matters because software was part of a broader interest in how things operate, rather than an isolated attraction to screens.
Dynatrace’s careers material describes him as one of its original developers and says he has led research and development from the beginning. That is a different kind of continuity from retaining a ceremonial founder title. The product changed, but the engineering function remained central to his job. The company dates a major cloud-native platform reinvention to 2015, followed by the introduction of Grail in 2022.
The campus project extends that relationship with the city. Its published plans connect engineering with the local technology community and research at JKU. Greifeneder describes the campus as a meeting place for domestic and international IT, AI and security communities. The aspiration gives the location a job beyond supplying office space: make it easier for people working on related problems to encounter one another.
A map for the next operator
To understand what kept changing in his technical assignment, consider the difference between a collection of readings and an account of events. A system can produce abundant measurements while leaving the relationship between them obscure. If one service slows and another fails, a chart for each does not automatically explain their connection. Someone, or something, has to work out the dependencies.
Greifeneder has argued that observability must include the business and the user’s experience. Metrics, traces and logs are useful, but treating their collection as the entire task misses the reason anyone wanted them. His concern is with the larger picture: what happened, how the system’s parts affected one another, and what the consequence was for the person trying to use it.
He had outlined the machinery of that ambition in July 2025, describing the third-generation platform through three connected activities: assembling knowledge, reasoning about it and taking action. Grail and Smartscape supplied data and relationships; causal, predictive and generative AI contributed different ways of interpreting them. Automation then provided a route from a decision to a change in the system.
For a reader outside software, the sequence is helpful. First establish what is going on. Then decide what it means. Then do something whose effects can be checked. Greifeneder’s argument puts those activities inside a continuing loop. The system’s response becomes information for the next decision. It is a substantial extension of diagnosis, and its usefulness rests on how faithfully each step represents the world it is meant to manage.
His 2026 writing carries that argument into AI. In January, he introduced Dynatrace Intelligence as an agentic operations system combining deterministic AI, contextual analytics and agentic AI. In his account, humans establish goals and policies while agents use live information to investigate and act. The ambition is autonomous operations with an account of why an action makes sense.
By May, he was emphasizing that human-led and agent-led software environments would coexist. Enterprises would need visibility across both as they changed how software was developed and operated. That is a less theatrical picture of transition than an overnight replacement of everyone at the keyboard. Existing systems and emerging workflows have to share the same operational world.
His August article turned to the data architecture underneath those agents. He argued for an AI lakehouse that connects live operational information with meaning and relationships. Greifeneder positions Grail within that foundation. His premise is that an agent needs relevant context prepared for it, including what depends on what, if it is to act reliably.
The connection to the founding question is easy to see. An automated operator inherits the difficulty that once frustrated a human engineer. It needs to understand the route from an action to its effects. Making actions faster increases the value of that understanding. A very efficient misunderstanding remains an expensive thing to put in charge.
The workshop and the water
Greifeneder’s practical interests continue outside enterprise software. He has described projects with his children involving soldering, coding, 3D milling and printing. One undertaking was a guitar built from scratch. His early software work included monitoring sack-weaving machines in the Dominican Republic. The objects vary considerably; the willingness to get involved with their workings is familiar.
Then there are foiling catamarans. “Sailing foiling catamarans is a huge passion of mine,” he has said. When the wind is absent, sanding hydrofoils supplies another task. It is an excellent arrangement for someone reluctant to run out of things to adjust. Even a quiet afternoon can apparently be given an engineering specification.
He also advises startups and supports academic research. His current company biography places that work alongside the effort to advance autonomous, intelligent digital systems. After a career spent making software behavior more legible, the next question is how much of its operation software itself can undertake. Somewhere, a customer is about to click. Greifeneder still wants to know what follows.