Dynatrace (NYSE: DT) today announced extended capabilities for the observability of customers’ GenAI initiatives. These advancements provide teams with access to comprehensive insights into their AI applications to drive reliability, performance, security and compliance. With this visibility, organisations now have clarity into their AI initiatives and can understand their return on investment (ROI).

IDC forecasts that worldwide spending on AI-supporting technologies will surpass US $749 billion by 2028. Investing in this technology is imperative for enterprises seeking to gain a competitive edge. Equally important is the ability to observe these technologies to help ensure they’re implemented optimally and securely, add business value, and lend themselves to higher degrees of automation.

Dynatrace is enabling its enterprise customers to achieve this through a series of industry-first platform advancements, including:

• Enhanced LLM Model Analytics: In addition to monitoring standard KPIs such as input and output errors, response times, and token consumption, the predictive capabilities of Dynatrace Davis AI® detect changes in usage behaviour to predict and forecast cost changes associated with LLM usage. This helps teams understand model performance and optimisation opportunities, including how they can better manage costs and control ROI.

• LLM Input and Output Guardrails: Dynatrace safeguards the quality of AI application input and output to help build trust in AI. This enables customers to recognise model hallucinations, identify attempts at LLM misuse such as malicious prompt injection, prevent Personally Identifiable Information (PII) leakage, and detect toxic language.

• Multi-model Tracing: Dynatrace maps dependencies between multiple LLMs that work in concert with Retrieval Augmented Generation (RAG) pipelines or agentic frameworks to provide end-to-end observability of the entire system, not just the component parts. This gives teams the insight to verify that dependencies are interacting seamlessly so they can deliver optimal end-user experiences.

• Responsible AI Integrations: Dynatrace helps organisations with AI governance by tracking every input and output without sampling to provide an audit trail of monitoring and observability, including documenting what training data was used for a given model. Through Dynatrace Grail, all data can be queried in real time and stored for future reference. 

Customer Value in AI Observability

“An increasing number of enterprise and mid-market merchants are leveraging artificial intelligence to empower their end-to-end payment strategy. At FreedomPay our challenge is to support merchants who look to seamlessly connect back-end systems with customer facing channels,” said Chris Kronenthal, President at FreedomPay. “Dynatrace enables FreedomPay to adopt advanced analytics and intelligence to solve complex data problems for our customers. This is required to deliver a world-class solution to improve compliance, security, and overall costs.” 

Learning from Industry Experts

According to Stephen Elliot, IDC Group Vice President, I&O and Cloud Operations, “AI represents a transformational change in technology and business.  Data and LLM observability are essential requirements for building and deploying reliable, trustworthy, and cost-effective AI solutions. These capabilities provide the critical insights needed to ensure that models perform as expected, meet business objectives, and deliver outcomes to the organisation.  Technology leaders benefit from monitoring their data quality in real time. They get visibility and detection of anomalies, inconsistencies, and drifts to help ensure that the data used to train and run the models is accurate, complete, and representative of the real world. This should be part of a broader observability strategy to deliver a robust, trustworthy platform that empowers organisations to prevent and resolve issues like hallucinations, drift, and performance degradation.”

The Dynatrace Vision

“We see a large portion of our global customer base moving their AI applications into production. AI Observability is key for ROI, governance, and explainability,” said Alois Reitbauer, Chief Technology Strategist at Dynatrace. “Dynatrace delivers AI-powered observability with real-time insights which enables data and systems to work together effortlessly. At Perform 2025, we’re excited to showcase how we’re leveraging these capabilities to power new possibilities for our customers, highlighting the transformative innovation they’re driving through the ability to effectively understand and optimise their AI deployments.”

Dynatrace supports customers today with its Observability for AI solutions. At Perform 2025, AI Observability will be the spotlight of the MainStage session on February 4 at 4PM PST.

To register to attend Perform 2025 in Las Vegas, February 3-5, in person or virtually, please visit: https://dynatrace.com/perform/

To learn more about the current list of technologies and integrations supported by Dynatrace AI and LLM Observability, visit the Dynatrace website