Reporting and surveys

Overview

Record when your AI Agent transitions to Messaging. Use that date when you compare conversation, campaign, and survey results.

Messaging does not count every legacy SDK interaction in the same way. A reporting change does not always indicate changed end-user behavior.

Survey completion rules depend on the survey configuration. Selecting a score without submitting does not always produce a completed response.

Do not assume that optional follow-up questions make every response complete. Review the current survey table before comparing results.

Use the Messaging reporting differences as the detailed comparison.

Area	Validation focus
Conversation counts	Compare opened and engaged conversations, especially when the previous integration used preload.
Surveys	Test each survey type, required questions, submission, and early closure.
Campaigns	Check when a Proactive message renders and when campaign events appear.
Transcripts	Review attachment links, reconnect events, retries, and survey entries.
Personalization	Verify retained metadata against the current field and value limits.
Links	Compare carousel and other link-click events used by your reporting.

Validate Messaging with examples from your production configuration.

To compare results:

Record the transition date and the integration that each test uses.
Run a representative conversation for each configured survey and human support path.
Compare the transcript with the saved reporting record.
Check your exports and webhook consumers against their documented event contracts.
Annotate the transition in external reporting before comparing trends.

Share unexpected differences with your Ada team. Include the integration, survey configuration, and conversation identifier.