camelAI auto‑builds dashboards from SQL
camelAI announced a tool that connects Postgres, Snowflake, BigQuery and MongoDB to an AI layer that accepts plain‑English queries, auto‑generates SQL and scaffolds dashboards with live URLs. (x.com) The pitch is speed‑to‑insight: fewer manual queries, but it raises the usual questions about governed metrics and who controls the semantic layer. (x.com)
Most business data still lives behind a gate called Structured Query Language, which is the database language analysts use to ask questions like “show me revenue by week.” camelAI says its new workflow lets a person type that question in plain English, then have the system write the query, build charts, and publish a shareable page on its own. (camelai.com 1) (camelai.com 2) The company says it can connect directly to PostgreSQL, Snowflake, BigQuery, and MongoDB, which covers both classic relational databases and document-style databases used by many software teams. Its product pages also say the agent reads the schema, writes code, and deploys results to a live Uniform Resource Locator, which is the web address someone else can open. (camelai.com 1) (camelai.com 2) That pitch lands because the normal route from question to dashboard is slow. A manager asks for “monthly signups since launch,” an analyst inspects tables, writes Structured Query Language, fixes joins, chooses a chart, and then rebuilds the same logic again inside a business intelligence tool. (camelai.com) (docs.getdbt.com) camelAI is selling the shortcut. Its own pages describe “one conversation” that turns a database connection into notebooks, dashboards, reports, and admin panels, with separate landing pages for PostgreSQL, Snowflake, and BigQuery promising production tools in minutes. (camelai.com 1) (camelai.com 2) (camelai.com 3) (camelai.com 4) This is part of a bigger shift in analytics software from hand-built dashboards to “agentic” analytics, where an artificial intelligence system does the query writing and chart assembly for you. Cube, Databricks, Snowflake, and Rill are all pushing versions of the same idea: let people ask questions in natural language, but keep the answers tied to governed business definitions underneath. (cube.dev) (databricks.com) (snowflake.com) (rilldata.com) The hard part is not drawing the chart. The hard part is deciding what a word like “revenue,” “active customer,” or “conversion” means, because two teams can use the same label and still count different things. (docs.getdbt.com) (atscale.com) That is why data teams talk about a semantic layer, which is the shared rulebook that tells every tool how to calculate a metric. dbt Labs says its semantic layer centralizes metric definitions so downstream tools use the same numbers, and Databricks now makes a similar case for governed metrics that work across dashboards, notebooks, and artificial intelligence agents. (docs.getdbt.com) (databricks.com) If camelAI sits above the database and generates Structured Query Language on the fly, the obvious question is who owns that rulebook. If the company is just a faster front end over a trusted metric layer, it could save analysts hours; if it invents logic ad hoc from table names and prompts, two people could get two different answers from the same warehouse. (docs.getdbt.com) (rilldata.com) (cube.dev) There is also a permissions problem hiding inside the demo. camelAI’s documentation says users define credential fields and connect services directly, which means companies still have to decide which tables the agent can read, which queries it may run, and whether a published live page can expose sensitive data. (camelai.com) (camelai.com) So the real product is not “artificial intelligence that makes charts.” It is a bet that companies will trade some manual dashboard work for a system that can move from schema to answer to web page in one step, as long as the trusted definitions of the business stay under someone’s control. (camelai.com) (docs.getdbt.com)
