Improving Data Visibility for a Modern Banking Platform with dbx

The Challenge

Say someone asks: "How many new users signed up in the last 30 days?"

For an experienced analyst, the SQL is obvious:

The query isn't the challenge. The workflow is — open an editor, find the table, verify column names, write it, run it, validate, share. Now multiply that by dozens of questions a day. Most data teams lose hours to requests that aren't technically hard, just endlessly repetitive.

Our Solution

SQL isn't going away — the layer above it is changing

Whenever AI enters a technical conversation, the same question shows up: "Will AI replace SQL?"

Almost certainly not. SQL is expressive, standardized, and deeply wired into modern data systems. What AI changes is the layer above SQL. Instead of hand-writing every query, you describe what you want and let AI draft it — the database equivalent of code completion. You still review the output and validate the logic; you just stop writing boilerplate from scratch.

That shift moves teams from syntax to intent. Instead of starting with GROUP BY, JOIN, and HAVING, you start with:

• "Show revenue trends by month."
• "Which customers haven't logged in recently?"
• "What were our top-performing products last quarter?"

The system translates intent into an executable query. The interesting part isn't that SQL gets generated — it's how much faster teams move from question to answer.

Faster exploration, not just faster queries

Analysis rarely stops at the first query. You run one, notice something odd, ask a follow-up, run another, repeat. That loop is where the time goes.

AI removes friction from the loop by keeping the conversation in natural language:

→Show revenue by month.

→Now filter for enterprise customers.

→Compare that with last year.

Instead of rebuilding queries each time, you stay focused on the data itself.

The Results

The shift is essentially one box in the pipeline:

Same pipeline — only the interface layer changes, from writing SQL to describing intent.

SQL still exists. Databases still exist. Developers still review and optimize. But the interface layer is changing — and just like GitHub Copilot changed how developers write code, AI assistants are changing how teams interact with data.

AI isn't replacing database professionals. It's removing the repetitive work that sits between a question and an answer. The most interesting part of this shift isn't automated SQL generation — it's making data more accessible without giving up the power and flexibility developers expect.

The future of database tooling may not be about writing better queries. It may be about writing fewer of them.