You asked, we delivered.

What's new: You can now set a default Board and Source for each Voice AI agent directly in your settings.

Why this matters: Until now, Voice AI tickets used your Messenger default settings. This meant calls often landed in the wrong queue and were hard to filter in your PSA. No more.

With this update, you can:

How it works:

Works across all supported PSA platforms.

Your feedback made this happen:

"We would want to be able to distinguish between chat-created tickets and Thread Voice–created tickets within ConnectWise. So, I have two different sources." — Kurt, WBM

"Based on what they choose here, you might want them to go through a different experience and create it on a different board." — Wil

"Source is the main thing I need... I can filter out and be like, hey, these are voicemails, these are active calls." — Cole, CCS Technology Group

Questions or feedback? Drop them below.

Inbox Teams provide a structured way to they can access. Teams let you group people once and assign Views at the team level, reducing the need to manually update View access as your organization changes.

As teams grow and change, managing View access one person at a time doesn’t scale. Inbox Teams reduce the ongoing setup and maintenance work for Admins by letting access be managed at the team level—making it easier to keep the right people working from the right Views as your organization evolves.

Learn more about Inbox Teams.

Real-world schedules aren’t perfectly clean—technicians often juggle overlapping work, and last-minute changes. These updates make Planner more flexible and realistic, giving teams a more wholistic view of their day without blocking progress, while still clearly surfacing conflicts when they matter.

We've improved how Voice AI transfers calls to your technicians.

When the AI transfers a call, the technician now hears a prompt and presses 1 to accept before the caller is connected.

Previously, the AI could mistake IVR systems, voicemail prompts, or recording disclosures for a live person — leading to awkward situations where the AI started talking to an automated message. Now, the handoff only continues once a real person confirms they're ready.

This should make technician handoffs more reliable, especially for teams using phone systems with IVRs or call recording announcements.

Questions or feedback? Let us know in the comments!

We've just released a new feature that gives you control over how your AI agent pronounces specific words.

Voice → Agents tab → Set up (on any agent) → General tab → Custom Pronunciations (under Communication Style)

You can map any word to exactly how you want it pronounced. This is especially useful for company names, acronyms, or unique brand terms. For example, you could map "ABC" to "A B C Technologies" or "Acme" to "ACK-mee." These mappings apply to everything your agent says.

We recommend using this for unique identifiers only—things like your company name or product abbreviations. Avoid mapping common words like "Technologies" or "Solutions," since those would get replaced everywhere in conversation.

A preview feature so you can hear the pronunciation before saving.

Let us know how it works for you!

Technicians and service desk managers can now give direct feedback on the Triage Agent’s work, either on the overall thread or on specific messages. This helps us continuously improve the agent’s accuracy, behavior, and usefulness over time.

This is our first step toward deeper visibility and continuous ML-driven improvement. More analytics and admin-facing insights coming soon.

.

Ask Magic can now search up to 200 past threads for both contact-level and company-wide questions (400 total!), up from the previous limit of 10.

This gives technicians far deeper historical context when investigating patterns, recurring issues, or past user behavior.

The experience remains safe and controlled: Ask Magic will never surface all 200 threads at once. Guardrails ensure only relevant examples are shown, while the LLM uses the broader context to generate more accurate insights.

No UI changes, this enhancement is entirely behind the scenes and available now.

.

Technicians can now upload images and PDFs directly in Ask Magic while working from Inbox. This makes it easier to reference files as you Ask Magic for help during a conversation with a contact.

.

The Triage Agent now uses stored contact intelligence more effectively to shape its clarifying questions. This allows the agent to reduce noise, skip redundant queries, and focus on what actually matters for faster, more accurate triage.

The agent is increasingly leveraging memory even in very short conversations:

➡️ Strong upward progression: the more conversation context available, the more intelligently the agent uses stored memory.

This results in sharper questions, fewer unnecessary follow-ups, and more context-aware triage.

After months of beta testing and micro-adjustments, we’re releasing the  to General Availability for Pro Partners.

The foundation of a stronger, ITIL-aligned triage model built for the agentic era.

We’re excited to announce the , a feature that uses to build a contact memory. This gives technicians and the triage agent valuable insights for a

Contact Intelligence captures key details about users - including their - helping your team better understand who they’re supporting.

Admins now have full control over where Magic Agents operate with the new Client Access page, live today for Triage Agent and Reminder Agent.

🔹 Choose or exclude specific clients, boards, or sources for each agent.

🔹 The setup mirrors what’s already available in AI Essential, ensuring a consistent experience across all Magic features.

🔹 Update permissions directly from each Agent’s settings, no extra Flows or steps needed.

This highly requested update gives admins precision and flexibility in managing access, so every Agent works exactly where it should. ⚙️💫

is officially live for all partners, introducing AI-powered mood tracking and actionable insights across every thread.

.

Sentiment scoring turns every interaction into measurable feedback, helping teams understand tone, act faster, and strengthen customer relationships.

now , not just attachments.

Previously, this limitation meant the agent was missing a lot of valuable context from emails, leading to a poor triage agent experience. With this update, even small details shared via pasted screenshots are now recognized, making for a smoother, more accurate triage agent conversations.

A small improvement with a big impact.