Skip to content

Agents and Athena

Ask the question. Read the query. Decide.

Questions about the numbers no longer wait for an IT ticket. Ask Athena inside Muvia, or the AI assistant you already use. Every answer comes with the query that produced it, so you can check it.

Book a demoFor IT: how it works

Try a question

AthenaSales project

Which areas lost the most orders compared with last quarter?

Three areas are down: North-East (−18%), Centre (−9%) and South (−4%). North-West and Islands are growing.

Query run

SELECT area,
  countIf(quarter = '2026-Q3') AS q3,
  countIf(quarter = '2026-Q2') AS q2
FROM orders
GROUP BY area
ORDER BY q3 - q2

Source: ERP (SQL Server)

Illustrative example

Two ways to ask, the same data.

Inside Muvia there is Athena. Outside, the assistant your company has already chosen. They answer on the same data, with the same permissions.

Inside Muvia

Athena

  • Answers on the project's data and shows the query
  • Writes, fixes and explains SQL
  • Builds the chart to put on the dashboard
  • Searches the documents you upload: contracts, manuals, procedures
  • Prepares the steps of a flow and the code of a function

Where you already work

Your AI assistant

  • Claude, ChatGPT, Microsoft Copilot and Cursor
  • Connects with your Muvia login, not a pasted password
  • Reads projects, alarms, flows and analyses; with your permission it runs SQL
  • Launches a flow or creates an analysis only after you confirm
  • The administrator decides who can connect it, and can revoke it

The questions people really ask.

  • Management

    Which customers ordered less than last year, and by how much?

  • Management control

    Explain how this margin is calculated, line by line.

  • Sales

    Make me a chart of revenue by agent for the last six months.

  • Purchasing

    What does the framework contract say about delivery penalties?

  • Analysts

    This query is slow and returns the wrong total. Fix it.

  • Data scientists

    Write me the function that computes the 7-day moving average.

An answer you can check.

You can see how it got there
Every answer carries the query it ran. Whoever knows SQL checks it; whoever doesn't asks for it to be explained.
It sees only what you see
It works with your permissions and on the projects you can access. Queries are read-only.
Switched on one piece at a time
Each AI capability is a separate module: the company turns on only the ones it wants.
Said plainly
Your data is in Milan; questions to the AI go to third-party language model providers.

Bring us the question that has no answer today.

In a demo we put it to Athena on your data, and you see the query that answers.

Book a demoFor IT