Closed queries
Each kind of question is a reviewed query. The AI chooses which one to use and with which filters, and cannot make data up.
“How much have we sold to this customer this year?” Today that question ends up in a report or an email to whoever knows how to pull the figure. An assistant connected to the ERP answers it on the spot. The AI doesn’t write SQL: it chooses between queries that are already approved, so every figure comes from a reviewed calculation.
Each kind of question is a reviewed query. The AI chooses which one to use and with which filters, and cannot make data up.
The assistant cannot change anything in your ERP. It connects with a read-only user.
How each figure is calculated in your company, written down and editable, so the assistant answers the way your team would.
The answer in whatever format is needed, in Spanish, English or Chinese.
What gets asked and which query was used, to review answers and add the ones that are missing.
Python, OpenAI or Claude, SQL Server, A3 ERP, Microsoft Fabric
What changes compared with hiring an agency or off-the-shelf software
Power BIPower BI for factoriesPower BI not refreshingSlow Power BIPower BI gateway errorAutomation with n8nEmail orders into the ERPWebsites and appsServices