Skip to main content

Objective

Create structured data tables that your AI agents can query using natural language. Perfect for customer databases, inventory tracking, order management and more.

Access

Sidebar -> Data Tables Path: /app/{tenant}/sql-tables

Roles with access

  • owner, admin, agent

Limits per plan

PlanTablesMaximum rows
Free1100
Starter35,000
Growth520,000
Business1050,000
EnterpriseUnlimitedUnlimited

Create a table

  1. Press Create Table
  2. Complete the form
  3. Press Create Table

Form fields

FieldMandatoryFormatExampleNote
Table Name (Internal)Yeslowercase + _customer_ordersInternal identifier
Display NameYestextCustomer OrdersVisible in UI
DescriptionNotextCustomer ordersHelp agents
SchemaYesfield listorder_id: TEXTDefine structure
Allowed OperationsYesmultiple selectionSELECT, INSERTLimit operations

Available field types

TypeDescriptionExample
TEXTFree text”John Smith”, “ABC123”
NUMBERWhole numbers or decimals100, 99.99
BOOLEANTrue/Falsetrue, false
DATEDate only2024-01-15
TIMESTAMPDate and time2024-01-15T14:30:00

View and manage data

Path: /app/{tenant}/sql-tables/{id}

Available Features

  • Add Row: Add a new record
  • Edit Row: Edit an existing record
  • Delete Row: Delete a record
  • Natural Language Query: Query with AI
  • Pagination: Navigate between data pages

Add a row (Add Row)

The form generates a field for each column of the schema. Notes by type:
  • NUMBER: Accepts decimal numbers (ex: 99.99)
  • BOOLEAN: Yes/No selector
  • DATE: Date picker
  • TIMESTAMP: Date and time selector

Natural Language Queries

This is the most powerful feature of Data Tables. Write a question in English or Spanish and the agent translates it into a SQL query.

Queries that work very well

TypeSample question
Count records”How many clients do I have?"
"How many orders are there?”
Sum (SUM)“What is the total sales?"
"What is the total revenue?”
Average (AVG)“What is the average order value?"
"What is the average order value?”
Maximum (MAX)“What is the highest sale?"
"What is the highest sale amount?”
Minimum (MIN)“What is the smallest order?"
"What is the minimum order?”
Filter by value”Show orders over $100"
"Find customers from Mexico”
Filter by date”Orders for January 2024"
"Orders after January 1st, 2024”
Sort results”Show the last 10 orders sorted by date"
"Show orders sorted by amount descending”
Combinations”What is the total sales for orders over $50?"
"Average price of products added this month”

Practical examples per use case

For an Orders table

"How many orders are there in total?"
"What is the total sales amount?"
"Show orders over $500"
"Orders from the last month sorted by date"
"What is the largest order?"

For a Customers table

"How many customers do we have?"
"Customers from Mexico"
"Customers registered after January 1st"
"Show the last 5 customers added"

For a Products table

"How many products are there?"
"What is the average price?"
"Products with a price above $100"
"Most expensive product"
"Cheapest product"

Features not supported (yet)

OperationStatusNote
GROUP BYIn development”Sales by region” does not work yet
JOINNot supportedTables cannot be related
OR conditionsNot supportedJust AND to combine filters
SubqueriesNot supportedSimple queries only

Technical notes

  • COUNT works with any number of records (100% accuracy)
  • SUM, AVG, MAX, MIN are calculated on a maximum of 500 records
  • Results are limited to 100 rows per query
  • Dates must be in ISO format (YYYY-MM-DD)

Assign tables to agents

For an agent to query a table:
  1. Go to Agents -> select the agent
  2. In the Tools section, activate Data Tables
  3. Select the tables that the agent can query
  4. Save changes
The agent will now be able to answer questions about those tables in chat, email, or voice conversations.

Good practices

  1. Descriptive names: Use clear names for tables and fields
  2. Fields in English: Keep field names in English for better compatibility
  3. Limit operations: Only enable INSERT/UPDATE/DELETE if necessary
  4. Useful descriptions: Add descriptions that help the agent understand the data
  5. Correct Types: Use NUMBER for numeric values, DATE for dates

Troubleshooting

”No results found”

  • Verify that there is data in the table
  • Check the spelling of the searched values
  • Try a simpler query first

”Query failed”

  • Avoid using OR in your questions
  • Do not ask to group by category (GROUP BY)
  • Make sure the fields exist in the schema

The agent does not respond about the table

  • Verify that the table is assigned to the agent
  • Confirm that the agent has the “Data Tables” tool enabled
  • Check that “Allowed Operations” includes SELECT

Screenshot

Data Tables List