Case study · Aerospace manufacturer · Business intelligence
A chat assistant that turns plain-English business questions into database queries against the ERP and returns the results, with the query shown for checking.
AI assistantBusiness intelligenceERP dataNatural language
The problem
How we solved it
How it works
Type the question the way you would say it out loud.
The assistant uses its map of the ERP to write the query.
The query runs with row and time limits.
Results appear as a table, with the SQL shown for checking.
Download to a spreadsheet, or ask a follow-up question.
Screens
One screen. The answer, the query behind it and the export button are all in the same place.
SELECT TOP 5 s.name AS supplier, COUNT(*) AS open_orders,
SUM(po.total_amount) AS open_value
FROM purchase_orders po WITH (NOLOCK)
JOIN suppliers s WITH (NOLOCK) ON s.id = po.supplier_id
WHERE po.status = 'Open'
GROUP BY s.name
ORDER BY open_value DESC;
| Supplier | Open orders | Open value |
|---|---|---|
| Northwind Fasteners | 14 | $88,420 |
| Harbor Metals | 9 | $74,150 |
| Delta Precision | 11 | $61,905 |
| Summit Coatings | 6 | $40,310 |
| Ridgeline Supply | 7 | $33,760 |
Illustrative screens rebuilt with sample data. No customer, supplier or employee information is shown.
What changed
People get numbers when they need them, not after a queue of report requests.
The query is shown with every result, so it can be verified, saved and reused.
Business rules live in one place, so the same question gets the same answer for everyone.
Trend and period questions use snapshots, which keeps production transactions responsive.
Analysts spend their time on new analysis instead of one-off lookups.
Use a hosted model, or run one locally when data must stay on the internal network.
Tell us what is slowing your team down. We will tell you honestly whether and how we can help.