Skip to content
ai-agent.review — AI Agent Reviews, home
← All news

Product updates

Ada adds handoff reasons to conversation exports and MCP

Vendor-reportedChecked: [release]

Editorial attribution: Gosia Madej

Prepared with AI assistance. Editorial attribution does not imply personal product testing or manual review of every story.

Published: · Event date:

Ada’s October 15 release exposes handoff reason codes and explanations through its Data Export API and MCP tools. Support teams can investigate why automation attempted a transfer, while checking its outcome separately.

Vendor-reportedChecked: [release]

What changes for users

Before comparing handoff causes over time, keep a separate unknown-code bucket and check the reporting denominator. Reconcile repeated exports by the documented target and timestamp rules instead of appending every row. Compare attempt outcomes with conversation escalation status before treating a cause count as successful human service.

Editorial assessmentChecked: [export][guide][filters]

A diagnostic tag hangs from a conversation card beside a route with a visible gap before a doorway marked by a person.
AI-assisted editorial illustration of a recorded handoff reason and a separate transfer outcome; not an Ada interface. · AI-assisted original vector illustration for AI Agent Reviews

The release describes a handoffs array in v2 conversation exports, reason fields in MCP get_conversation and a HANDOFFREASON filter. Ada’s guide dates reason-code collection to October 7; the October 15 announcement concerns access to those reasons, not a new start date for all collection.

Vendor-reportedChecked: [release][guide]

The export reference specifies the latest attempt per handoff target, including outcome_status. It warns that a later export can contain fewer entries. For merging exports, Ada documents matching on handoff_id, falling back to id, and retaining the entry with the later date_created; ties use the later export.

Vendor-reportedChecked: [export]

A reason can be recorded even when a transfer fails. The guide distinguishes a missing code, null, from the recorded value unclassified. Counts by recorded reason omit handoffs with a null code; ISNOT also returns conversations with no handoff. These are important boundaries for an escalation dashboard.

Vendor-reportedChecked: [guide][filters]

Checked against sources

AI-assisted source check

How this was checked

AI-assisted checks read the full dated release, handoff guide and export/MCP references. The hub and individual release now agree on October 15; the export and filter references now include the previously missing fields. All evidence is vendor documentation.

Editorial assessmentChecked: [release][export][guide][filters]

No live API call, tenant rollout test or source screenshot was obtained. These fields describe attempts and documented classifications; they do not independently establish successful human contact or the accuracy of an assigned cause.

Editorial assessmentChecked: [release][guide]

Source evidence

Original source excerpt. This shows what the source says; it does not establish product performance.

even if the attempt fails

Read original source [guide]

Sources

Related agents