Ask in the web app
Open a project's Chat and ask in plain language. Cassis answers with the SQL and the context it used.
Sign in at app.getcassis.com, choose a project, and open Chat.
1. A straight answer
The context defines revenue, so Cassis answers straight away and shows the SQL and the objects it used.
What was revenue in 2025?
Revenue in 2025 was €4.22M: item price on delivered orders, shipping excluded.
sql: SELECT SUM(oi.PRICE) AS revenue_eur
FROM STALLORA.ORDER_ITEMS oi
JOIN STALLORA.ORDERS o ON oi.ORDER_ID = o.ORDER_ID
WHERE o.ORDER_STATUS = 'delivered'
AND YEAR(o.ORDER_PURCHASE_TIMESTAMP) = 2025
objects_used: business_volume, STALLORA.ORDER_ITEMS, STALLORA.ORDERS
FROM STALLORA.ORDER_ITEMS oi
JOIN STALLORA.ORDERS o ON oi.ORDER_ID = o.ORDER_ID
WHERE o.ORDER_STATUS = 'delivered'
AND YEAR(o.ORDER_PURCHASE_TIMESTAMP) = 2025
objects_used: business_volume, STALLORA.ORDER_ITEMS, STALLORA.ORDERS
Show SQL 👍 / 👎 Add as test case
2. Clarification, then the answer
Nothing in the context defines an active customer, so Cassis asks first. Pick an option or reply, and it runs. Your choice holds for this conversation only.
How many active customers do we have?
One thing to confirm
What makes a customer active? Nothing in the context defines it.
A delivered order in the last 90 days Any order in the last 12 months
A delivered order in the last 90 days
3,912 customers had at least one delivered order in the last 90 days, counted once each across their orders.
3. Follow-up in the same chat
A follow-up reuses the earlier question, so you do not restate it.
Now by product category.
2025 revenue by category: health_beauty €512K, furniture_decor €448K, sports_leisure €406K, electronics €371K, and 66 more.
objects_used: business_volume, STALLORA.ORDER_ITEMS, STALLORA.ORDERS, STALLORA.PRODUCTS
Read the answer
- Clarification
- When essential business information is missing, Cassis asks first, with choices when it has them. Answer and the question runs; your answer holds for this conversation only.
- SQL
- Show SQL expands the exact statement behind the answer.
- Objects used
- The tables, joins and metrics the answer used, so you can spot an unexpected one.
- Feedback
- Rate the conversation up or down and give the required reason.
- Add as test case
- Editors and Admins can save the question and its SQL as an eval case. Have someone validate the SQL first.
Correct a wrong answer
Reply in the same chat, for example “Revenue should exclude shipping.” Cassis uses the correction for the rest of that conversation. It does not change the shared context: like a rating, it is evidence your data team sees in Review once Cassis analyzes that chat.
A conversation keeps the context it started with. Once the fix is published, ask again in a new conversation to check it.