The data is in the database. The answers are not.
Many companies keep their most important data in their own database: orders from the website, activity in the product, records from a system built in-house, or a data warehouse in BigQuery. The data is complete and up to date. The problem is who can ask it.
Only people who write SQL can. Everyone else sends a request to IT or to the one data person, and waits. Simple questions — how many new customers this month, which region is growing — sit in the same queue as real projects.
How it works
- Create a dedicated database user. You decide its permissions and which tables or views it can see. Cloud databases and BigQuery need nothing installed; an on-premise database connects through a secure agent, with no inbound ports to open.
- WizBI learns your data. Tables, columns and the business terms behind them — what "active customer" or "net sales" means in your company — in Hebrew and English.
- Your team asks in the chat. In plain language, no SQL. WizBI works out the query, runs it and answers with a number, a table, a chart or an Excel file.
- It keeps working when nobody asks. A weekly report on schedule, or an alert when a figure crosses a line you set.
For IT: the safety model
The usual worry about tools that let you talk to your database is simple: what stops it from changing or exposing data it should not? Here is the model, point by point.
| Concern | How WizBI handles it |
|---|---|
| Access | A database user you create, with the permissions you choose. You decide which tables or views it can see. |
| Changes to data | Only what you agree to. Any automation that changes data is defined with you on request, runs under the user's permissions, and every action is logged. |
| Network | Cloud databases and BigQuery: nothing to install. On-premise: a secure agent, no inbound ports. |
| Isolation | Each organisation's data is isolated and encrypted in transit and at rest. |
| Who sees what | Permissions by organisation, group and user — the same on WhatsApp, Teams, Slack and the web. |
| Audit | Every question is logged and can be reviewed. |
| AI models | They run inside Google Cloud under enterprise terms. Your data is not used to train them. |
WizBI cannot see more than the database user you create. If that user cannot read a table, neither can WizBI.
WIZBI AI LTD is certified to ISO/IEC 27001:2022, and WizBI runs on Google Cloud.
More than one database
A database rarely holds the whole story. Orders are in the database, invoices are in the ERP, targets are in a spreadsheet, projects are on monday.com. WizBI connects to all of them — Priority, SAP, monday.com, Google Sheets, Excel and CSV files, BigQuery, SQL databases and more — so one question can draw on several sources.
The same answers are also available inside AI tools your team already uses, such as Claude or Microsoft Copilot, through MCP — with the same permissions. See WizBI over MCP.
Natural language to SQL (text to SQL): which tool for whom
If you are an analyst who writes SQL every day, a SQL editor — with or without an AI helper that drafts queries — is the right tool. You want to see the query, edit it and own it.
WizBI is for the people around you: managers, finance, sales and operations who need the answer, not the query. It answers in the chat they already use, knows your business terms, combines the database with your other systems and sends reports on schedule. No BI team or data skills needed.