Ask a business question.
Get an evidence-backed answer — no SQL required.
QueryDash is an autonomous investigation agent that turns a plain-language business question into a multi-step, evidence-based analysis of your operational data — without your team writing a single query.
An investigator, not just a query box
QueryDash doesn't just answer one query — it plans, investigates, checks its own work, and explains what it found.
Natural-Language Investigation
Business users type a question in plain English. No SQL, no data-team ticket, no waiting.
Autonomous Multi-Step Reasoning
The agent plans its own investigation — deciding what to check next based on what it just found.
Dynamic Schema Discovery
Explores your actual tables, columns and business terminology — no hardcoded queries to maintain.
Evidence-Backed Findings
Every answer comes with the tables, metrics and visualizations that support it — never a guess.
Read-Only by Design
Every generated query is validated server-side before it runs. Writes and schema changes are never possible.
Real-Time Investigation Streaming
Watch the agent work step by step instead of waiting on a black box for the final answer.
From a question to an answer, in four steps
The same flow, every time — auditable, explainable, and safe to run against production data.
Describe the Objective
A user states a business question in plain language — "Why did revenue drop in the west region?"
Agent Plans the Investigation
QueryDash determines what data is needed and discovers the relevant tables and fields.
Controlled, Read-Only Analysis
It generates and validates SQL, runs it safely, evaluates the results, and drills deeper where needed.
Evidence-Backed Summary
Findings, key metrics, charts and a recommendation are returned — with the evidence behind them.
Enterprise infrastructure, managed end to end
No self-hosted models, no GPU infrastructure to manage — just AWS managed services doing the heavy lifting.
- Amazon Bedrock with Amazon Nova Pro powers agent reasoning and tool use.
- Amazon ECS on AWS Fargate runs the production application behind an Application Load Balancer.
- Amazon RDS (MySQL) and Amazon Aurora Serverless v2 (PostgreSQL) connect as investigation data sources.
- Amazon DynamoDB holds short-term conversation and investigation context.
- Amazon S3 & CloudFront deliver the frontend securely at scale.
Technology Stack
AI-generated queries never touch your database directly
Every statement the agent produces passes through a server-side validation layer before it executes — the model never gets a direct line to production data.
Read-only enforcement
Only SELECT, WITH, SHOW, DESCRIBE and EXPLAIN are permitted. Writes and schema changes are rejected outright.
Row-limited queries
Unbounded queries are automatically capped to protect performance and prevent runaway results.
Isolated production data
Databases sit in private subnets, never directly exposed to the application's public surface.
Managed credentials
Secrets are stored in AWS Secrets Manager; service access is governed by AWS IAM.
Built for teams who are tired of waiting on data
Self-service for business users
Non-technical teams get answers directly, without filing a request and waiting on a SQL specialist.
Answers dashboards can't
Ad-hoc, one-off questions get investigated properly — not deflected because "there's no report for that."
Safe by default
Full investigative power over your data, with zero risk of an AI-generated write or schema change.
Any operational database, any recurring question
See QueryDash investigate your own data
Bring a real business question to the demo — watch the agent plan, investigate and explain its findings live.