Run the AI Guide analysis

The AI Guide is STYD’s schema understanding layer. It analyzes your database structure so the AI can better understand tables, columns, and relationships.

The first analysis

When you connect a PostgreSQL or MySQL database, your first analysis is included while your workspace still has an AI Guide setup left (a workspace without a plan has 1; plans have more every month: see Pricing). It starts by itself, for databases of up to 3,000 columns.

If no free setup is left, or the database is larger, STYD shows the price first (about 1 prompt for every 500 columns, at least 1) and runs the analysis only after an owner or admin chooses Analyse. “Later” leaves it for later. The database already works in Chat before the analysis finishes; the analysis adds the AI Guide and the database summary, which make answers more accurate.

Odoo connections have no analysis of this kind: STYD reads Odoo itself, and the Odoo AI Guide is rebuilt when you re-sync.

When to run it again

Run the analysis again when your database structure changes. Once the first analysis is done, every new one costs prompts (about 1 for every 500 columns, at least 1), and STYD asks you to confirm the price before it starts. Regenerating the database summary costs 1 prompt.

New columns were added
The AI Guide needs to analyze new columns before STYD can use them reliably.
Column meaning changed
If a column is reused for a different business purpose, edit its description; the analysis never overwrites a description you saved.
New important tables were added
Run the analysis again when new business tables become available.

Sample values matter

The AI Guide uses table and column structure, but also sample values when available. These samples help STYD understand what a column represents in practice.

If a new column has no data yet, the analysis may detect the column name but miss useful sample values. Add representative data before running it again when possible.

What you can review and edit

The analysis creates AI-drafted descriptions for your tables and columns. These descriptions help STYD understand your database, but they can also be reviewed and improved by your team.

Table descriptions
Review and edit AI-drafted explanations of what each table represents.
Column descriptions
Review and edit AI-drafted explanations of what each column means.
Table visibility
Control which tables are available or hidden for specific users.
Column metadata
Review detected type, sample values, PII, URL fields, and soft-delete indicators.

Column details

For each column, the AI Guide can display useful metadata that helps both users and the AI understand how the column should be used.

Column type
Sample values
AI-drafted description
PII detection
URL detection
Soft-delete flag

Why the AI Guide improves answers

Without schema context, an AI can easily choose the wrong table, misunderstand a column, or produce a query that does not match your business logic.

Better table selection
Better column understanding
More accurate SQL generation
Clearer answers from complex schemas