Using Tables Inside Agents
AI Tables are not just for data storage — they are the foundation of how your AI agents reason, remember, and act.
When connected to an agent, tables can be used to read, write, and update information in real time during conversations.
This allows your AI agents to go beyond chat — they can manage structured data, create records, and perform actions automatically inside your workspace.
How agents use tables
When an AI agent is connected to a table, it can:
Retrieve data (for example: “Show me all open support tickets.”)
Create new records (for example: “Add a new lead with name Ana López.”)
Update existing data (for example: “Mark ticket #102 as resolved.”)
Cross-check information between multiple tables.
📸 Screenshot of an agent conversation showing AI reading data from a Support Tickets table
Agents use the table structure to understand what fields exist and how to map user inputs to them.
This makes interactions more natural and connected to your internal data.
Storing new information automatically
When a user shares new information during a chat, the agent can save it directly into a table.
For example:
User: “My company is called NovaTech and I need a quote.”
→ The agent creates a new record in the Leads table with the user’s name, company, and context.
📸 Screenshot showing an agent automatically creating a new record from a chat interaction
This helps your organization capture information passively, without forms or manual data entry.
Querying data in real time
Agents can also use tables to retrieve live data and answer questions accurately.
Examples:
“What orders are still pending delivery?”
“Who is the assigned agent for ticket #203?”
“How many new leads were added this week?”
📸 Screenshot showing agent response generated from a live table query (Orders table with pending records)
The agent uses the table as its knowledge source and can access only the data you allow, ensuring full control and security.
Updating existing data
AI agents can modify or enrich records based on conversation context.
For instance:
Updating a ticket’s status after resolving an issue.
Adding a follow-up date after confirming a sales call.
Changing a lead priority based on sentiment or intent.
📸 Screenshot showing agent updating an existing table record during conversation
You can define which fields agents can edit and whether updates require validation or approval through workflows.
Using relations between tables
If your tables are related (for example, Customers → Orders), agents can navigate those links naturally.
For instance:
“Show me all orders from customer Lucas García.”
“Add a new invoice for this customer.”
📸 Diagram showing an agent accessing connected tables through relations (Customers → Orders → Invoices)
Relations give your agents full context — they can access all relevant information without switching between systems.
Combining agents, tables, and workflows
The real power comes when you combine these three components:
Agent collects data from a user.
Table stores it and structures it.
Workflow reacts and performs actions automatically.
For example:
A customer requests a quote → Agent creates a Lead → Workflow sends a notification to Sales.
A support request is resolved → Agent updates Ticket → Workflow sends a feedback form.
📸 Screenshot showing agent, table, and workflow connected in one automated flow
This integration allows your agents to act as part of your operational team — not just chatbots, but fully functional digital assistants.
Best practices
Clearly name your tables and fields to help the AI map information correctly.
Use Select and Relation fields for structured and consistent data.
Combine tables with workflows for real-time automations.
Define which tables and fields agents can access for security and clarity.
What’s next
Now that your agents can use tables, explore how to take it further:
📸 Screenshot of a workspace dashboard showing agents and tables interacting in real time
Tip:
AI Tables turn your agents into active collaborators.
They don’t just answer questions — they manage, update, and trigger actions, making your workspace truly intelligent.