Agentic AI Platform Change Log

MCP / tool calling / API

Ada has removed its scripted proactive campaigns feature and deprecated the associated campaign actions within the Embed2 Web SDK. While existing integrations will continue to load, the deprecated actions will no longer have any effect and will log a console warning.

Engineering and CX teams using the Embed2 SDK for proactive messaging must update their implementations, as these scripted triggers will no longer function.

Security / enterprise

Ada introduced a suite of security and compliance capabilities for managing AI agents, including an audit log that tracks configuration changes across the dashboard, API, and MCP interfaces. The update also adds bulk end-user deletion for GDPR compliance and MCP connection management to monitor and revoke authorized AI-assistant connections.

Enterprise security teams can now maintain strict oversight over how AI agents are modified and by whom, especially as external AI assistants gain the ability to propose changes via MCP. The bulk deletion tool also provides a verifiable mechanism to meet strict data privacy and compliance obligations.

Integrations

Ada has released a native integration enabling its AI Agent to directly answer Instagram direct messages. Connected entirely from the dashboard with a Meta login, the integration handles inbound and outbound text and media without requiring third-party middleware.

Organizations can now deploy their Ada agent on Instagram to automatically resolve customer inquiries on a major social channel. Bypassing external middleware lowers technical overhead and simplifies the architecture of an enterprise's social media support operations.

Agent capability

Ada has introduced availability rules for Glossary terms, enabling administrators to restrict when specific definitions are applied based on user variables. A term whose rule does not match the current conversation context is skipped. These rules can also be managed programmatically via MCP.

Support teams can now tailor their agent's vocabulary and definitions strictly to specific customer segments or conversational contexts. This improves the accuracy of the agent's responses and ensures specialized terms are only triggered when relevant.

Observability / auditability

Glossary usage is now surfaced directly in the Ada Conversation View through transcript events that indicate when the AI Agent applies glossary terms. Administrators can review translation modes and definitions used per turn, as well as filter conversations and Analytics reports by specific terms.

This enhanced visibility allows AI trainers to verify that the agent is properly adhering to brand terminology in real conversations. It drastically simplifies debugging and evaluating the effectiveness of the glossary in steering agent responses.

Security / enterprise

Ada released the Bulk End-User Deletion API to programmatically erase personal data associated with specific users across its systems. The API accepts up to 1000 identifiers per request via email, external ID, or stored variables, automatically propagating the erasure downstream across all data stores.

This update allows enterprise compliance and security teams to fully automate data privacy operations in accordance with GDPR and CCPA requirements. Programmatic control removes the manual overhead of processing deletion requests and ensures comprehensive data removal across all connected support records.