Natural Language Queries

Grepp lets you ask questions about your data the same way you would ask a colleague. No SQL required — just describe what you want to know and Grepp handles the rest.

How to Ask Questions

Grepp understands a wide range of question styles. You can be specific or high-level; Grepp will ask a clarifying question if it needs more information.

Good examples to get started:

• "Show me total revenue by month for the past year"
• "Which customers placed the most orders last quarter?"
• "Compare signups this week vs last week"
• "What are the top 10 products by units sold?"
• "How many open issues were created in the last 30 days?"

Grepp reads your connected data source schemas automatically, so you never need to name tables or columns — just describe the information you are looking for.

Supported Query Types

Grepp can handle a broad range of analytical questions across your data:

Aggregations and summaries — totals, averages, counts, sums broken down by any dimension (region, category, status, user, etc.)

Filters and conditions — questions scoped to a specific time range, status, customer segment, or any other attribute in your data

Comparisons — side-by-side views across time periods, segments, or data sources (for example, this month vs last month, or plan A vs plan B)

Time series — trends over time grouped by day, week, month, quarter, or year

Rankings — top N or bottom N results sorted by any metric

Multi-source questions — when you have multiple integrations connected, Grepp can query them independently and present results together

What Happens When You Ask

When you send a message, Grepp:

1. Understands your intent and identifies the relevant data source
2. Generates and runs a query against your connected integration
3. Interprets the results and explains them in plain language
4. Builds a chart or table on your dashboard if a visualization makes sense

If Grepp has already fetched similar data earlier in the conversation, it reuses those results rather than re-querying — so follow-up questions are fast.

Tips for Better Results

Mention the time range. Questions like "last 30 days," "this quarter," or "since January" give Grepp clear scope and produce more accurate results.

Name the metric you care about. "Revenue," "signups," "open tickets" — the more specific you are about what you want to measure, the more precise the result.

Ask follow-ups naturally. After Grepp builds a chart, you can say things like "now break that down by region" or "show the same thing but for last year" and Grepp will update the result in context.

Be approximate when unsure. If you are not sure how your data is structured, describe what you want in plain terms. Grepp will explore the schema and find the best match — or ask you a clarifying question if it needs one.

Destructive Actions

For actions that permanently change your dashboard — such as removing a chart or deleting a dashboard — Grepp will always ask for your confirmation before proceeding. No data or dashboard content is ever deleted without an explicit "yes" from you.