Toby AI TruePilot: AI Observability That Works Inside Your Current View
When engineers investigate production issues, the evidence is already on the screen — a dashboard, a log query, a time range, a trace, a RUM page, a monitor, or a filtered list of hosts.
The problem is that AI assistance usually lives somewhere else. An engineer finds a group of 500 errors in a log viewer, copies a sample into a separate chat window, explains the filters, describes the time range, waits for a response — and then gets an answer that's too generic, because the assistant never had the real observability context.
Toby AI TruePilot is built for a different workflow. It works as a contextual AI assistant inside the TrueWatch page you're already viewing, using the current dashboard, filters, chart definitions, time range, and page state as part of the conversation.
Why Context Should Stay With the Investigation
The core difference between a standalone AI assistant and an AI observability tool like Toby AI TruePilot is context.
A standalone assistant waits for you to describe the situation. Toby AI TruePilot starts from the current page: the dashboard being viewed, the monitor page, the log filter, the time window, the chart configuration. You ask a question in natural language without manually carrying the evidence into another tool.
This doesn't mean Toby AI TruePilot has unlimited context or authority. It means the assistant starts from the same operational view as the engineer — which makes the first answer more relevant and cuts the copy-paste loop that slows investigations down.
How Does Toby AI TruePilot Generate Dashboards?
When you want to create or modify a dashboard, Toby TruePilot turns an analysis goal into a first draft. For example:
Analyze API error rate over the last seven days and show the top five services.TruePilot generates chart definitions, titles, descriptions, and an initial layout as an AI dashboard, ready for review. You still check the result, adjust the query, and decide whether it belongs in a shared dashboard.
Root Cause Analysis AI for Events and Incidents
For alerts and incidents, Toby AI TruePilot collects related signals — logs, traces, containers, metrics, hosts — and produces an analysis draft explaining what changed and where to look next.
The valuable part isn't a final answer. It's a structured first pass across several observability signals:
- Affected service or endpoint
- Related trace or log patterns
- Container or host context
- Metric changes in the same time window
- Possible downstream dependencies
- Evidence that supports the hypothesis
Treat this as investigation support, not a verdict. Root cause still needs to be confirmed against evidence and ownership before anyone acts on it.
Generating DQL From Plain Language
The bare term "DQL" has the top two SERP results occupied by Dynatrace's own "Dynatrace Query Language" official documentation (also abbreviated as DQL), with the fourth and sixth positions taken by OpenSearch and Wikipedia's generic definitions — the three-letter abbreviation has been claimed by the most direct competitor. This is not a problem that keyword ranking alone can solve, but at minimum, expanding the term on first mention ensures that readers and search engines immediately understand this refers to TrueWatch's own query language, creating clear differentiation from Dynatrace's identically named abbreviation.
Show error logs from the order service in the last hour.TruePilot generates a query across relevant observability data sources and runs a syntax check. You review the dataset, filters, groupings, and time range before saving or reusing it.
Documentation Retrieval Without Leaving the Page
Configuration questions come up constantly during investigation — a parameter name, a setup step, a troubleshooting path. Searching for it interrupts the flow.
TruePilot retrieves relevant TrueWatch documentation and returns configuration guidance with links, keeping you on the current task instead of sending you into a separate documentation search.
Page-Level Analysis
TruePilot can also analyze the current page context directly — a dashboard, a RUM page, a trace view, or another data page. Depending on the page, it can summarize visible signals, explain frontend performance patterns, inspect trace error distribution, or flag the data dimensions worth a closer look.
What's Coming Next: In-Page Workflows
The following workflows are planned for future iterations and should be confirmed as generally available before you rely on them in production.
Logs — ask about the current error set. Filter the log viewer with source:nginx AND status:500 over the last hour, then ask Toby TruePilot: "What is causing these errors?" TruePilot drafts an explanation from the current filter and sample logs — for example, upstream connection timeouts concentrated on a payment API path — without you pasting log samples into a separate chat window.
Dashboards — add a comparison line. Viewing a line chart for today's order volume, ask: "Add a comparison line for the same period last week." TruePilot reads the current chart definition and drafts a time_shift(7d) query so the comparison shows up in the same chart.
Monitors — draft an alert rule. Viewing a host monitoring page filtered to CPU usage above 80%, ask: "Create an alert for this host." TruePilot reads the current host ID and metric context, then drafts a rule such as: trigger when CPU usage on web-server-03 stays above 90% for five minutes. The important word is draft — alert rules affect operations and need review before they're enabled.
Why This Matters
Toby AI TruePilot lowers the operational cost of context switching:
- User intent is grounded in the current page
- Analysis happens inside the workflow instead of outside it
- DQL, dashboard, documentation, and page analysis all work from visible context
- Less experienced users can start from natural language, while experienced users keep control of the final query, chart, or rule
This is part of the broader TrueWatch AI direction. Toby AI TruePilot helps you understand and work with the current observability context. Toby AI Agents goes further — bringing role-based agents, permission boundaries, evidence trails, approval flows, and governed actions into production operations.
TruePilot helps you think inside the current page. Agents help teams coordinate work across production workflows. People stay responsible for judgment, approval, and high-impact decisions.
FAQ
Q: What is AI observability (also called artificial intelligence observability)? A: AI observability is the use of AI to help engineers understand and act on production data — logs, traces, metrics, dashboards — from inside the tools they already use, instead of a separate chat window. Toby AI TruePilot applies this directly inside TrueWatch, using your current page as context instead of asking you to describe it from scratch.
Q: What is Toby AI TruePilot? A: Toby AI TruePilot is TrueWatch's AI observability assistant. It works inside the page you're already viewing — a dashboard, log filter, monitor, or trace view — using that context to answer questions, draft dashboards, generate DQL queries, and analyze incidents.
Q: How is this different from a general-purpose AI assistant? A: A general-purpose assistant needs you to describe your situation from scratch — filters, time range, the data you're looking at. Toby TruePilot starts from the current page state, so the first answer is grounded in the actual dashboard or query you're already working with.
Q: Can Toby AI TruePilot create alerts or dashboards automatically? A: TruePilot drafts dashboards, DQL queries, and alert rules — it doesn't publish or enable them automatically. Every draft is reviewed by the user before it's saved, shared, or turned on.
Q: Does Toby AI TruePilot replace root cause analysis? A: No. TruePilot provides a structured first pass across logs, traces, metrics, and host context to support investigation. Confirming root cause against evidence and ownership is still the team's responsibility.
Q: What's the difference between Toby AI TruePilot and Toby AI Agents? A: Toby AI TruePilot helps you think and work inside the current page. Toby AI Agents goes further, adding role-based agents, permission boundaries, evidence trails, and governed actions for coordinating work across production workflows.
Next Step
Open TrueWatch, go to the page where you're already investigating, and ask Toby AI TruePilot a question grounded in the current view.
