Why BI is not enough
A manager has to know SQL or wait for an analyst's report; every question goes through the analyst's queue; and cloud AI services see the content of the queries and the results along the way. A specialized agent removes all three problems at once.
Four steps of the process
- Voice or text: the agent recognizes the speech and extracts the business meaning of the request.
- Access check: the user, their role and the allowed data area are determined.
- SQL query: only to the allowed PostgreSQL views; writing is forbidden.
- Answer with the source: the result, the period, the calculation method and the table to double-check.
Who benefits
- Managers: operational metrics by voice without BI and analysts.
- Analysts: a quick check of metrics without writing SQL.
- Finance teams: revenue, margin, accounts receivable — by voice or text.
- Production: warehouse, manufacturing and delivery data in real time.
Boundaries
The answer depends on the quality of the data model and the dictionary of business metrics. The agent has no write access and does not bypass corporate access policies: access to sensitive data is restricted through a role-based access model. Details — on the subagent's page.
Questions and answers
Can the agent change the data in the database?
No: only read-only access to the allowed views; writing and modification are forbidden at the architecture level.
Do I need to know SQL?
No: the question is asked in natural language — in text or voice, and the agent builds the query itself.
Is the data sent to an external service?
In local mode — no; cloud mode is possible only by the user's explicit choice.
Can the consultant modify the data?
No: the connection is read-only by architecture — only the allowed PostgreSQL views.
Is voice supported?
Yes: a voice question is recognized and turned into the same query to the database.
What does the answer include?
The data from the database, an explanation of the calculation and a link to the source table.
How does the consultant know the company's terms?
A dictionary of business metrics is connected: metrics are calculated according to the definitions adopted in the company.
Which database is supported?
PostgreSQL; access is configured to the allowed views.
Where is the consultant deployed?
A local perimeter for confidentiality or cloud models for scalability — by choice.
What if the answer is not in the database?
The consultant reports that there is no data instead of guessing: every conclusion is linked to the source table.