Assembled AI Agents

We would like to introduce a pre-engagement message on the chat interface. Specifically, we need the capability to display custom text on the homepage before the user selects the 'Ask us a question' prompt to help set expectations or provide immediate guidance

We would like to request a feature for multiple email agent. Our current workflow involves several inbound email streams that require unique handling. Ideally, we need the capability to configure and update independent agents for each email channel to ensure specialized responses for different query types.

We have identified a significant number of instances where accurate support responses were blocked from delivery by automated safety filters. To optimize our workflow, we are requesting the ability to toggle these safety checks on or off at our discretion.

Since integrating Assembled for chat/messaging, our agents have lost visibility into their CSAT scores. Two specific gaps: Agent Scorecard — The "Average CSAT Score" section only pulls from Zendesk native CSAT. Because our tickets are scored via Assembled CSAT (Sunshine API channel), this field doesn't populate at all. Agents can no longer track their own scores from the scorecard. Ticket-level drill-down — Agents previously could open a messaging ticket and see their CSAT score + client feedback inline. This no longer works for Assembled/Sunshine API tickets. Request: Surface Assembled CSAT data in the Agent Scorecard's CSAT section, and allow agents to click into a score to view the associated ticket and client comment — parity with what native Zendesk CSAT previously offered. This was raised directly by frontline agents in our Q1 roundtable and affects the majority of our CSAT volume (~99% comes through Assembled/Sunshine API).

When using Assist with Zendesk, it would be incredibly beneficial to be able to use the smart fields and data to update Zendesk ticket fields without first requiring a workflow to run on the interaction. The data (UID, OID, URLs, etc) are being populated into an internal comment at handoff, but cannot populate a ticket field without a workflow running. Having this information in a ticket field allows for things such as: Advanced reporting on which URLs are driving the most contacts Better routing to ensure the right team member is receiving the ticket Increased visibility on device type And more!

Customers currently have no way to automatically detect whether images attached to Zendesk tickets (or other support channels) are AI-generated versus authentic photos/screenshots. This creates challenges for: Fraud prevention teams dealing with synthetic documents Trust & safety teams moderating user-generated content Support teams needing to verify the authenticity of customer-provided evidence Requested Feature: Add capability to analyze attached images and flag or score them based on likelihood of being AI-generated or manipulated, rather than authentic photos. Use Cases: Detecting fraudulent identity documents, receipts, or proof-of-purchase images Identifying synthetic product photos in marketplace/platform abuse scenarios Flagging deepfakes or manipulated images in user content moderation Risk scoring tickets that contain potentially synthetic evidence

Ability to export report data found in Assembled

The ability to set operating hours within Assist is critical; otherwise, Chat interactions can be handed off when no live Agents are available. For customers with both WFM and Assist products, the operating hours should be reflected in both platforms.

Request for related tickets for manual bulk solves (without merging) when customers contact multiple times about the same topic and there are "open" tickets.

Requesting to add zendesk ticket form as a property when creating a trigger.