Curated notes on the latest Sundial releases, workflow upgrades, and product improvements.
Workspace administrators can give each user group access to the appropriate warehouse account, limit it to selected databases and schemas, or revoke access to a connector. Sundial also filters the tables, semantic models, and measures shown to users and the AI analyst.
Routines let teams schedule recurring Sundial analyses daily, weekly, or monthly at a chosen local time. Important checks can run automatically and deliver concise updates to the right Slack channels.
Teams can delete Data Apps they no longer need directly from the Library or by asking Sundial. Deleted Apps disappear from normal product and agent workflows, making recurring analysis easier to keep organized.
Playground lets data teams test changes to dbt models, semantic models, playbooks, organizational context, and evals on an isolated branch before merging. Teams can run evals against the branch and catch regressions before changes reach production.
Teams can use the metrics and semantic models they already manage in dbt MetricFlow directly in Sundial, bringing trusted business definitions into analysis without changing their existing dbt workflow.
Teams can stop an in-progress analysis, add new context, and steer what Sundial should do next. This makes longer investigations easier to guide when the original question changes or new information becomes available.
Sundial can connect to Google Workspace sources such as Gmail, Drive, Calendar, and Docs, helping teams bring the operating context behind a metric or question into the same analysis workflow. Admins can manage these connected sources at the workspace level.
Teams can use Sundial to update the semantic layer, playbooks, and other organizational context without manually editing the underlying files. This creates a more direct path from feedback to the context that guides future analysis.
Sundial now surfaces upstream and downstream lineage for semantic models, measures, and dimensions in a graph view. Teams can follow how a metric is defined and connected, making analytical context easier to inspect and trust.
Build an App by asking Sundial for the view you need, from KPI cards and charts to filters and custom visuals. Apps stay current when you open them, and the Library gives teams one place to find, share, favorite, version, restore, and continue editing their Apps and Artifacts.
Connect approved GitHub repositories so Sundial can answer analysis questions with the codebase context behind them. Teams can investigate instrumentation, events, metrics, and implementation details alongside their data without switching between tools.
Sundial can turn analysis into reusable, refreshable app-like outputs that teams can revisit as the underlying data changes. This helps recurring analysis live beyond the first conversation without becoming a static screenshot or one-off export.
Sundial agents can now use connected apps while answering questions, so teams can combine warehouse metrics with the docs, issues, and discussions that explain what is happening. Connected apps help bring operating context into analysis without copy-pasting it into every conversation.
Sundial data-modeling workflows now integrate more directly with GitHub-backed semantic context. That gives teams a clearer path from model changes to reviewed, durable analytics context that fits the development workflows they already trust.
Quick Chart helps teams explore measures with filters, breakdowns, timeframe controls, and shareable chart URLs. It gives teams a faster path from a metric question to a focused visual view when they do not need a full investigation.
Teams can mention Sundial in Slack, ask analytics questions in natural language, and get an answer where the conversation is already happening. When the question needs charts, tables, artifacts, or follow-ups, the answer can link back into the web app for deeper analysis.
Sundial now gives teams clearer progress visibility while it works through an analysis. The goal is to help users understand the shape of the task and stay oriented during longer-running work without exposing raw chain-of-thought.
Sundial's AI analyst now runs on a more resilient foundation designed for real analysis work. It can use tools iteratively, stream progress, load the right business context on demand, and keep longer-running work moving even when the browser connection changes.