Schema in, sensitive rows stay put
The LLM receives schema context such as tables, columns, and metadata so it can form the right query plan without needing your customer rows.
Zoqi is designed for teams that need conversational analytics without compromising privacy, production safety, or deployment control.
The LLM receives schema context such as tables, columns, and metadata so it can form the right query plan without needing your customer rows.
Use read-only credentials to ensure the system can inspect and query your data but never delete, update, or mutate operational records.
Deploy inside your own environment when you need tighter controls, internal networking, or enterprise procurement alignment.
A teammate asks a question in natural language.
Relevant schema, metadata, and semantic definitions are assembled.
The model generates the query using the grounded map, not raw datasets.
The query runs against your database using safe credentials.
The result is returned as an explanation, SQL, and visualization.