Feature Request: Support excluding replies to multiple handles in X (Twitter) API v2 to Control @-Mentions
1. What is the problem?
There is a critical disconnect between the agent's view in the Zendesk Agent Workspace and the final published reply on X (formerly Twitter).
When an agent replies to a tweet that contains multiple @-mentions, they can manually delete the unwanted mentions from the composer window.
The Zendesk UI correctly reflects these deletions, showing a clean reply addressed only to the intended user.
However, when the reply is published, the deleted @-mentions are still included in the live tweet on X. The agent's edits are completely ignored by the backend process that posts the reply.
We confirmed with Zendesk Support (Ticket details from chat on Aug 21, 2025 #13772666) that this is because the integration only uses the in_reply_to_tweet_id parameter and does not utilize the exclude_reply_user_ids parameter available in the X API v2.
2. Why is it a problem?
This discrepancy creates significant business challenges:
- Major Compliance & Legal Risk: For regulated industries like ours (FinTech), this is a major policy violation. Accidentally including government agencies (@FTC, @FBI), public figures, or unrelated brands in a customer service reply can create serious compliance incidents and brand damage. Our agents are trained to de-escalate and isolate conversations, but the tool is failing to execute their actions.
- Erosion of Trust in the Platform: The UI provides false confidence to our agents. They believe they are adhering to company policy, but Zendesk is publishing a tweet that violates it. This "What You See Is Not What You Get" behavior means we can no longer trust the composer for public social media interactions.
- Operational Inefficiency: To prevent compliance breaches, our agents must now leave Zendesk after every reply, manually check the live tweet on X to ensure it was posted correctly, and then delete it if Zendesk failed to remove the mentions. This adds significant overhead and slows down our response times.
3. How do you solve it today? (Workarounds & Their Limitations)
Zendesk Support suggested two workarounds, both of which are inadequate and result in a poor customer experience:
- Agent flags any ticket to specialist: When a ticket has multiple handles mentioned, our agents will flag to a specialist who then responds natively. The specialist is able to remove the extra @ mentions natively and then reply just to the original poster.
4. What is the ideal solution?
Zendesk should fully implement the functionality of the X API v2 for replies.
Specifically, when an agent deletes an @-mention from the reply composer in Agent Workspace, Zendesk should dynamically populate the reply.exclude_reply_user_ids array in the POST /2/tweets API call with the user IDs of the removed accounts.
The behavior should be simple and predictable: The content and @-mentions shown in the Zendesk composer right before sending must be identical to what is published on X.