Header background

Dynatrace expands Davis AI with Davis CoPilot, pioneering the first hypermodal AI platform for unified observability and security

Dynatrace is proud to announce the expansion of Davis® AI with Davis CoPilot™. With the addition of generative AI capabilities, Dynatrace is now the first hypermodal AI platform in the industry.

Hypermodal AI, which combines predictive AI, causal AI, and generative AI, boosts productivity across operations, security, development, and business teams.

This expansion of Davis AI complements the proven Dynatrace predictive AI model (for example, forecasting and anomalies) and our causal AI model (for example, determination of a problem’s root cause, security risks, user impact, and steering automation), which are at the core of the Dynatrace platform.

Davis CoPilot empowers users to effortlessly create queries, data dashboards, and data notebooks using natural language and provides coding suggestions for workflow automation, reflecting the unique attributes of each customer’s hybrid and multicloud ecosystem. It also simplifies and accelerates onboarding, configuration, and adoption of the Dynatrace platform.

Dynatrace hypermodal AI for unified observability and security
Davis® AI combines predictive AI, causal AI, and generative AI, making it the first hypermodal AI for observability and security. Predictive AI and causal AI provide deterministic answers and reliable automation, while the precise context additionally enriches generative AI for automatic or user-created prompts.

What is hypermodal AI, and why is it essential for reliable observability, security, and automation at scale?

Hypermodal AI intelligently combines multiple AI techniques—predictive AI, causal AI, and generative AI—helping organizations effectively solve BizDevSecOps use cases.

Dynatrace applies these techniques to the broadest set of modalities in the market, including the data types of metrics, traces, logs, behavior, topology, dependencies, events, and more, with unmatched precision for precise predictions, accurate determinations, and meaningful insights.

Davis AI transforms and augments data to enable more useful analysis, perform automatic tasks, and respond to user requests:

  • Predictive AI uses machine learning (ML) and statistical methods to recommend future actions based on data from the past. Dynatrace uses the various data types across metrics, logs, traces, behavior, events, and more in its Grail™ data lakehouse and causal dependencies from Dynatrace Smartscape® to provide continuous forecasting and anomaly prediction, including cloud application health, infrastructure needs, sales, and customer experience trends, seasonality, and other historical behaviors.
  • Causal AI processes observability, security, and business data in the context of causal dependencies from Dynatrace Smartscape topology to precisely determine the needle in the haystack in continuously and dynamically updated software services. It groups anomalies, pinpoints root causes, ranks security risks, enables precise attack investigation, and provides business impact assessments, all automatically. This AI also triggers automated remediation actions. It further enables teams to explore trends or patterns with built-in domain and topology context.
  • Generative AI drives productivity through AI-powered analytics and automation for all members of your organization. Davis CoPilot interprets natural language to create queries, dashboards, and notebooks and provides suggested code for automation workflows. It further simplifies access to best practices for observability and security use cases and answers “how-to” questions precisely. It also guides users who want to observe new technologies or apply advanced configurations.

The combination of AI techniques is vital for observability and security use cases

Generative AI is a transformative technology for delivering productivity gains. Observability, security, and business use cases raise additional challenges as they need precision and reproducibility.

Large language models (LLMs), which are the foundation of generative AIs, are neural networks: they learn, summarize, and generate content based on training data. When a user asks a question, generative AIs create an answer word by word. They predict the probability of the next word or sequence of words given the input prompt. They employ probabilistic sampling techniques and allow controlled randomness to diversify responses.

This means that the same prompt/question will provide different responses. Because this randomized, probabilistic approach is not rooted in precise causal data, a pure generative AI approach renders use cases that require precision impossible.

Davis AI combines AI techniques for precise and reliable outcomes:

  • Predictive AI and causal AI provide context to Davis CoPilot. They automatically enrich prompts with specific information, which provides better recommendations and precise, reproducible results.
  • Davis CoPilot generative AI doesn’t only react to user inputs. It can also be triggered automatically by predictive AI or causal AI events (for example, to recommend remediation actions automatically).
Davis AI enriches prompts with context, unlocking use cases that require precision and specificity
Davis AI enriches prompts with context, unlocking use cases that require precision and specificity.

Hypermodal AI unleashes exponential value: Step-by-step example

In this example, a user builds a Dynatrace dashboard for all business-critical services that will be at risk during Black Friday. The steps required to complete this task can be categorized as either predictive predictive AI icon, causal causal AI icon, or generative generative AI icon.

generative AI icon  Understand the meaning of questions.

causal AI icon  Identify all user sessions that contain conversion metrics (using Smartscape).

causal AI icon  Identify all services that are needed for these user journeys (topology using Smartscape).

predictive AI icon  Predict how these services will behave under a higher load based on historic data.

causal AI icon  Choose the services that are nearing their limits (topology metrics).

causal AI icon  Choose the services that caused problems in the past.

generative AI icon  Use input to generate a dashboard and queries.

generative AI icon  Determine if remediation workflows should be set up.

Dynatrace Davis® AI provides answers and automation, and boosts productivity for multifaceted use cases

Automatic root cause analysis

Davis AI automatically detects user-facing issues and assesses their impact on the business and affected users. Then, Davis uses context—such as topology, transaction, and code-level information—to identify the precise root cause of problems.

Davis CoPilot can provide recommended actions to remediate issues.

Automatic root cause analysis enables AIOps (or AI for IT operations) automation using the Dynatrace AutomationEngine.

Root cause with Davis CoPilot

Natural language queries

Dynatrace Query Language (DQL) is a powerful tool to explore data and discover patterns, identify anomalies and outliers, create statistical modeling, and more based on data stored in Dynatrace Grail. With this indexless approach, you can execute any query at any time.

Davis CoPilot translates natural-language questions into DQL queries, using causal AI for additional context, such as dependency information.

Generate DQL with Davis CoPilot

Auto-coded workflows

Davis CoPilot auto-generates code to make it easier to create workflows using natural language input.

Autoremediation workflows or automated integrations with ChatOps, DevOps, and ITSM tools have never been easier.

Auto-coded workflows with Davis CoPilot

Predictive operations

Davis AI enables forecasts with automatic anomaly prediction (for example, to autoscale resources). It can generate reports and take action automatically.

These actions can range from informing the respective team to automatically triggering orchestration actions.

Forecast series with Davis AI

Auto-generated quality checks

The Dynatrace Site Reliability Guardian allows development teams to define quality objectives in their code, which is validated throughout the delivery process before the code reaches production.

Predictive AI and causal AI apply machine learning, anomaly detection, and root cause analysis to make this easy. Davis CoPilot creates guardians for specific services with natural language input.

Auto-generated quality checks with Davis CoPilot

AI-powered application security

Davis AI not only assesses risks automatically; it also detects and blocks threats.

Davis CoPilot can recommend remediation strategies and simplify security analysis across all data by translating natural language into DQL queries that drive attack protection, security investigations, and forensics.

AI-powered application security with Davis CoPilot

Natural language to visual analytics

Powered by Grail data, Dynatrace provides visual tracking and analytics using dashboards and notebooks, leveraging dependency and topology data from Smartscape.

Davis CoPilot creates dashboards and Notebooks based on natural language input, fueled by causal AI.

Natural language to visual analytics with Davis CoPilot

AI-assisted onboarding and platform use

Whether you want to observe additional technologies, apply advanced configurations with one sentence, or leverage new capabilities and best practices, Davis CoPilot uses its custom-trained large LLM to boost productivity, ensure fast onboarding, and unlock AI for all members of an organization.

AI-assisted onboarding and platform use with Davis CoPilot

Davis AI with predictive AI and causal AI is generally available and used by all Dynatrace customers. Start your free trial now! Davis CoPilot™ will be available later in 2023 as a core technology within the Dynatrace platform. Find more information and interesting links here.