Your Data Team Shouldn't Be a Help Desk: Use Compass with Your Data

Your data team is doing the best they can, but demand for data is limitless. The problem is that getting answers from data means hunting down the right dashboard, figuring out if it's current, and translating what you see back into the question you asked. Most people give up or go with their intuition.

We built Compass to fix this. It's AI-powered data analysis that lives where you work. Ask a question in plain English, get an answer from your warehouse with charts and context. No dashboards. No tickets. No waiting.

And unlike other AI analytics tools, Compass works directly with your existing warehouse: Snowflake, BigQuery, Redshift, AWS Athena, Postgres, DataBricks, or MotherDuck. Your data stays where it is. Your governance stays intact. You just get faster answers.

Compass is available today at compass.dagster.io.

Fast exploratory analysis, where conversations happen

Mention @Compass in any channel where it's present and ask your question. Compass searches your governed context from your warehouse, writes SQL, and responds in the thread with explanations and visualizations.

"What is the likelihood a prospect will close if they attend a webinar?" Done in seconds, with a breakdown and trend chart.

"calculate the daily median, 75th percentile, and the daily max concurrent usage for our enterprise customers over the past 30 days.Then take the median of those 3 metrics, the 75th percentile of that. Give me a csv of the daily percentiles and the summary statistics" Compass does multi-step analysis and returns a csv for you.

Under the hood, Compass follows a sophisticated query lifecycle, from context search through dataset discovery to SQL generation and result processing. (Curious how it works? See How Compass Turns Questions Into Queries.)

What we've seen internally and with early users is that Compass actually improves the relationship between the data team and stakeholders. Instead of data teams being a bottleneck, they become teachers. When someone asks a question, the data team sees it, can jump in to add context or validate the approach, and learns what their stakeholders actually need. These interactions create better feedback loops. Data teams discover which datasets need better documentation, which business logic needs to be codified, and which questions come up repeatedly. The relationship shifts from transactional (ticket in, answer out) to collaborative (let's figure this out together).

Built for data teams who want control

You control which datasets Compass can access. You see every question and query. When something's off, you step in and help Compass learn.

The real power is in the context store. Your warehouse has the data, but you have the institutional knowledge, naming conventions, business logic, edge cases. Compass learns that "net sales" excludes refunds in your company, or that you always use billing_address_state rather than shipping_address. This context is what makes Compass accurate for your organization, not just technically correct.

When users hit something complex, they can bring you into the thread directly. You maintain governance without becoming a bottleneck.

Quick Setup, immediate value

Connect your warehouse, install the Slack app, you're done. Compass automatically scans your schemas and starts answering questions. No semantic layer to configure. No models to define. No documentation required.

A note on security: Your data never leaves your warehouse. Compass connects directly using your existing permissions and security model, generates SQL queries, and executes them with read-only access. We store the conversations and context you provide, but your actual database data stays in your infrastructure. You maintain complete control over what Compass can access. AI providers don't train on your queries; each customer gets complete workspace isolation, and all connections use TLS 1.2+ encryption with AES-256 for stored data.

It learns as your team uses it, getting smarter about your business with every question. The more context you provide about business logic and naming conventions, the more effectively Compass can generate accurate queries for your specific organization.

Answers you can trust

Compass is built for exploratory analysis, answers that inform decisions, and spark follow-ups. Every response includes the SQL and sufficient context to understand where the numbers originated. The analysis is also iterative and does several queries as it finds the best solution. 

When precision matters for board decks or commission calculations, hand it to your data team for a proper report. Compass handles the rough-cut questions where speed matters more than perfection, freeing your team for the work that requires deep expertise.

And because everything happens in Slack, context is preserved. Follow-up questions build on previous answers. Teams collaborate around data without switching tools.

What teams are using it for

RevOps checks pipeline health before meetings. PMs validate hypotheses about user behavior. Customer success looks up account metrics while on calls with clients. Engineers look at structured logs to identify product bottlenecks. 

One of the most powerful patterns we've seen: multi-user conversations. Someone starts exploring data with Compass, uncovers something interesting, and brings other teammates directly into the thread. The data team can review queries, add context, or suggest follow-ups, all without leaving Slack.

We've been using Compass internally at Dagster for months. Anil, our data analyst uses it to scale analysis without scaling the team, handling exploratory questions, prototyping dashboards, and building a centralized context store that makes every answer smarter. The common thread: questions where speed matters more than perfection.

Available today

Compass is live. If your team uses Slack and has a Snowflake, BigQuery, Redshift, AWS Athena, Postgres, DataBricks, or MotherDuck warehouse, you can connect it in five minutes at compass.dagster.io.

Compass is publicly available for everyone. Dozens of early adopters are already live and seeing amazing results. We’re shipping continuously and will continue to iterate over the next few months based on what we learn from teams in production.

Want to try it first? Join our Dagster Community Slack where we run a live demo instance. Ask questions, see how others are using it, and get a feel for Compass before connecting your own warehouse.

Coming soon:

• Proactive insights: Daily checks that surface trends and anomalies automatically
• Deeper governance: Granular permissions and audit trails
• Dagster+ integration: Native connection to your orchestration platform for richer context about freshness and lineage and streamlined data operations.
• Custom Implementation: If you wish to deploy Compass on Microsoft Teams, Discord, on prem or have any other unique considerations, reach out to us and we can work with you.

Why we built Compass

Dagster exists to make data engineers more productive. We built an orchestration and observability platform that treats data assets as first-class citizens and makes it possible to scale data platforms without losing your mind.

Compass extends that mission to the people who depend on the data those platforms produce. The same philosophy driving Dagster, respecting data infrastructure, working with your existing stack, prioritizing experience, applies here.

We designed Compass to work with your warehouse, not replace it. Whether you're running Snowflake, BigQuery, or Redshift, Compass connects directly to your existing infrastructure. No data movement, no vendor lock-in, no architectural changes required.

The data platform you've built powers the products your business runs on. Compass makes those products more useful by putting them where your team works.

Try Compass today at compass.dagster.io.

Compass connects directly to Snowflake, BigQuery, Redshift, AWS Athena, Postgres, DataBricks, or MotherDuck and additional warehouses coming soon. Slack integration requires workspace admin approval.