… 6 unchanged lines …
  
  ## Key facts
  
- - Available for **voice and live chat channels only** in the current preview release
+ - Available for **voice and messaging channels only** in the current preview release (earlier previews were limited to voice and live chat)
  - Playbooks consist of a **trigger event**, up to **10 conditions** (with up to 2 context variables each), and a set of **actions**
  - Supported trigger events include: conversation waiting in queue, conversation transferred
  - Supported actions include: increase priority (0–10,000 range), trigger overflow (route to voicemail, transfer to queue, redirect to number)
+ - Supported scenarios include dynamic prioritisation, overflow handling (now factoring wait time in queue and queue position alongside representative availability, sign-in status and operating hours), bullseye routing through user groups, and assigning to a preferred or previously engaged expert
+ - The preferred-expert scenario now covers **keeping the same expert for returning asynchronous chats**, checking the expert's presence status and active conversation limits before reassigning
+ - The prompt gallery now surfaces **six** popular prompts by default (up from three), with templates grouped by scenario
  - Playbooks have two statuses: **Draft** (editable, not active) and **Active** (live, changes require Save & Publish)
  - You cannot have two active playbooks for the **same scenario on the same queue** — this is a hard conflict
  - Different scenarios on the same queue (e.g. wait-time prioritisation + availability overflow) **can** run simultaneously
    … 28 unchanged lines …
  - Clients consistently underestimate the queue-mapping exercise. Before any playbook is written, you need a definitive list of which queues handle which conversation types and what their acceptable wait thresholds are. Without that, you end up with conflicting active playbooks across queues and a chaotic publish process.
  - The natural-language playbook authoring sounds great in demos, but in UAT you'll find that ambiguous prompts produce inconsistent condition matching. Spend time with the client defining precise trigger wording using the provided templates rather than freeform language.
  - The PAYG billing model catches procurement teams off-guard in scoping. Raise it as a dependency early — the Azure subscription approval can take weeks in regulated industries, and it's a hard blocker.
- - Preview features in D365 CC have a history of behaviour changes between preview and GA, particularly around condition logic and the context variables available per channel. Build this into your testing plan and don't make firm promises about specific variable support until GA.
+ - Preview features in D365 CC have a history of behaviour changes between preview and GA, particularly around condition logic and the context variables available per channel. The channel coverage has already broadened from live chat to all messaging during preview, so don't make firm promises about specific variable or channel support until GA.
  
  ---