Connect your tools

Living Context: custom connectors and field mapping

Living Context is a crucial component of the Unless AI agent that enables real-time access to customer data by connecting external systems such as CRMs, helpdesks, and other data sources. This capability allows the agent to understand what the customer is doing at the moment and tailor its responses accordingly. To achieve this, you configure connectors and field mappings that define how data flows from your systems into the agent’s context.

what is Living Context and why it matters

Living Context complements Living Knowledge by focusing on actions and current customer states rather than static knowledge. For example, it can inform the agent if a customer’s usage has dropped significantly or if they are at the limit of their subscription plan. This dynamic awareness enables more relevant and timely interactions, such as triggering renewal reminders or checking order statuses.

The agent does not invent actions but runs predefined procedures connected to your systems, ensuring consistent and authorized operations.

types of connectors and how to set them up

Unless supports three main types of connectors to integrate external data sources:

standard connectors

These are pre-built integrations for popular platforms, including:

• CRMs: Salesforce, HubSpot
• Helpdesks: Zendesk, Freshdesk, Jira Service Management, Zoho Desk, Xurrent
• Knowledge bases: Confluence, Google Drive, BookStack, Zenya, ClickHelp, Azure DevOps Wiki, WordPress
• Live chat: Zendesk Sunshine Conversations, Freshchat
• Data lakes: Visma Data Lake

To connect a standard system:

1. Go to Train > Living Context > Connectors > Add connector
2. Select the desired system from the list
3. Authenticate via OAuth
4. Choose the data scope you want to expose
5. Map the fields to be available to the agent

MCP servers

For custom or internal tools, you can connect your own MCP (Model Context Protocol) server. This allows the agent to read and act on proprietary data sources with the same audit trail and permissions as standard connectors.

custom connectors

If your system is not supported by standard connectors and MCP is not suitable, you can build a custom connector using the Unless connector framework. This requires developer involvement or assistance from the Unless team.

field mapping: controlling what the agent sees

Field mapping defines which data fields from your connected systems are exposed to the agent. This is important for privacy and relevance:

• Only fields included in the mapping are visible to the agent; all others remain in your source system.
• Sensitive fields can be marked as such and tokenized through the Privacy Vault, so the agent sees tokens instead of raw data unless directly responding to the data subject.
• You can inspect the connector to view sample data the agent will access during conversations.

This granular control helps ensure compliance with privacy policies and limits data exposure to what is necessary.

managing connector behavior and troubleshooting

Living Context connectors operate in real time, querying your systems at conversation time rather than relying on cached data. This ensures the agent always has the latest information.

If a connector is slow or unavailable, the agent gracefully degrades its responses, continuing to assist without the contextual data.

Common issues with connectors returning no data include:

• Loss of permissions for the user who authenticated the connector
• Field mappings pointing to fields that no longer exist
• Source system being unreachable

when to use custom connectors or MCP

Consider custom connectors or MCP servers when:

• Your system is not on the standard connector list
• You need the agent to access fields not included in default mappings
• Privacy reviews require limiting or customizing data exposure
• You are creating new Moments that require new or specialized data

Your account manager can help you scaffold custom connectors or guide you through MCP server setup.

summary

Living Context connectors and field mappings empower your Unless AI agent to act with up-to-date, relevant customer data from your existing systems. By carefully selecting connectors, mapping fields, and managing permissions, you ensure the agent delivers personalized, context-aware interactions while maintaining privacy and security. Whether using standard connectors, MCP servers, or custom-built integrations, Living Context is key to turning knowledge into action in your customer journeys.