Skip to main content

AI Agents

Build custom AI assistants that can query and interact with your connected data sources, platforms and processes​

AI Agents are custom AI assistants that understand your business data through connectors. Unlike generic chatbots, agents can answer questions and take actions using live data from your Salesforce CRM, ServiceNow tickets, JIRA projects, databases, and other connected systems.

Why use AI agents​

Business context awareness – Agents query your actual data sources to provide accurate, context-specific answers. Ask "What's the status of the Acme deal?" and the agent pulls real-time information from Salesforce.

Built-in tool calling – Agents automatically determine which connectors to query and what actions to take. Connect ServiceNow and the agent can search tickets, create cases, or update records as needed.

No prompt engineering required – Simply connect your data sources and the agent figures out how to use them. The underlying LLM handles tool selection and execution.

Multi-source reasoning – Agents can combine data from multiple connectors in a single response. Cross-reference a JIRA issue with Confluence documentation and Slack conversations to provide comprehensive answers.

What you can build​

Internal knowledge assistants – Connect your documentation, databases, and collaboration tools so employees can ask questions and get instant answers grounded in your actual data.

Customer support agents – Query CRM systems, support tickets, and knowledge bases to help customers with account-specific questions and issue resolution.

Workflow automation – Build agents that not only answer questions but take actions—creating tickets, updating records, sending notifications, and triggering business processes.

How agents work​

  1. Connect data sources – Add connectors to your application (Salesforce, ServiceNow, databases, etc.)
  2. Configure the agent – Choose which connectors the agent can access and set any behavioral parameters
  3. Deploy and use – The agent automatically queries connected systems and executes actions based on user requests

Agents use the underlying LLM's native tool-calling capabilities to determine when and how to use each connector, eliminating the need for complex prompt engineering or manual tool orchestration.

Learn more about connecting data sources by reading the connectors documentation, or get started building with the Agent Studio.

Next steps​

See the following guide to Create your first AI Agent.