Newly released frontier models from and are now available in .

If you pin your provider to OpenAI, you can now select

If you pin your provider to Grok/xAI, you can now select

All four of these new options performed very well in robust testing with various agents. To know what works best for you, we always recommend testing your agent.

If you're not sure which model to select, we always recommend selecting as the provider and as the model, which lets the system pick a model for you dynamically that the system determines will handle your current agent configuration well. will only ever run your agent with a "frontier" model (ie the most generally capable ones).

Ava got some significant intelligence improvements today — especially related to complex routing rules for booking — Ava can intuit and one-shot your super complex rules more consistently now. In addition to that, all Ava modes now reason more thoroughly before responding.

When generating booking rules and assignment rules, so you can visually see the 'map' of your rules. This should help spot check whether your rules are actually what you want instead of reading through a hundred lines of rules explained in text.

To check it out simply go into any agent that has booking or lead assignment in the Booking/Goal tab, and click on

You'll then see an interface like this which you can drag/zoom as needed to get the whole picture!

We recorded a video course which covers things like what you need to know on a day to day basis with how leads flow through the system, and overviews of major features. Check it out soon and work through it at your own pace!

It is at the top of the help home page

It can also be accessed in the section of the menu.

You can get to by clicking on the icon button in the main left hand menu in when you are logged into hub.brayv.ai

All AI Text Agents can now escalate a conversation to your team — outside of conditional tools. When an agent runs into a situation it isn't equipped to handle — a hostile contact, a legal or safety concern, an explicit request for a human, or something outside its instructions — it can decide to flag the conversation and alert your team with a specific reason explaining why.

Account admins will receive escalation alerts (in-app and email) by default. If you see one, that's an agent asking a human to take a look.

Escalated conversations show an "Escalated" badge on the Conversations page. Hover or tap it to see the AI's reason, and click Resolve once you've handled it. Replying to the contact manually also "resolves" it automatically.

Important: an escalation is an alert, not a handoff. The agent keeps working the conversation — answering questions, qualifying, booking, whatever it's configured to do — while your team is looped in.

What you can configure (Settings > Text AI > Escalation Alerts):

Who gets alerted — choose specific admins/staff, set email and/or SMS per person, and optionally alert the contact's assigned owner

Turn it off entirely — possible, but not recommended;

Two-strike hand-off (optional, off by default) — if an escalation goes unresolved and the agent has to escalate a second time, it sends a final hand-off message and stops responding for that contact

— add instructions or a Conditional Tool with the right actions for it, and the agent will handle it definitively (and quietly) next time. Escalations should get rarer as your agents get sharper.

Full details in the help article: (Help Center > AI Text Agents).

Your sequenced, scripted follow-ups can now speak more than one language. Pick a contact field (eg, create a custom field like "preferred language"), ensure your agent has field extractor tool to keep that field up to date, set a simple rule, and your agent automatically sends the right-language version of each follow-up — same timing, order, and cadence, just different text.

Generate a second-language version of your existing sequence with AI

Edit each language side by side, step by step

Falls back to your default messages whenever there's no match, so nothing changes for the rest of your leads

Great for accounts already have their agents converse in Spanish and route bookings for Spanish-speaking leads to the right reps — now your follow-ups match the conversation.

See new step by step guide in > >

— so the soonest appointment is never blocked by one teammate's calendar.

When you set up booking rules for your AI agents, a rule can now hand the calendar a pool of eligible users rather than one specific person. For example:

The calendar keeps using its own user-selection strategy, just limited to the pool you specified:

Maximize Availability — the AI offers the combined availability of everyone in the pool and books the least-loaded rep who's actually free. If one person is on vacation or fully booked, the contact still gets the next real opening from someone else in the group.

Optimize Distribution — picks the least-recently-booked rep within the pool - not the intended use case for this feature (because optimize distribution balances out to the least recently booked rep, keeping availability to a single user), but it will still work.

This is particularly useful when you want to book among a particular group of team members, but don't want to select a specific one before booking (which limits availability for your customers to that particular team members availability).

No more calendar bloat. Instead of creating a separate calendar for every combination of reps, put your team on one calendar and let rules slice them into the right pool per contact (by zip, service type, etc.).

Easier to maintain. Change who handles what by editing a rule — no juggling dozens of calendars to keep track of.

Better availability + automatic balancing. Contacts get the soonest real opening across the whole eligible group, and bookings spread evenly among them.

Works with text and voice AI agents, and with the manual "Let AI Select Calendar" button (you can let it auto-assign within the pool, or pin a specific person).

If a rule ever points only to people who aren't assigned to the chosen calendar, we stop the booking and flag it so you can fix the setup — rather than quietly booking the wrong person.

: see who was booked, why, and every change

Open any appointment to see exactly how it was assigned and a full timeline of every change — all in one place.

Each appointment now has an Appointment History panel with two parts.

For appointments booked by your AI agents, you'll see the user selection reasoning that was followed after the AI made the request to book the appointment.

Who was eligible, who got booked, and why (e.g., "fewest appointments in the last 30 days").

And — most usefully — why each other eligible rep wasn't chosen: they had more recent appointments, the time was outside their availability, it conflicted with an existing appointment or blocked time (including buffers), a travel rule, they weren't in the eligible pool, or they'd hit their appointment limit.

This answers the question we hear most — "Why didn't [rep] get booked? They haven't had an appointment in ages." — instantly, instead of cross-checking availability, blocked slots, buffers, and travel rules one at a time. It's the fastest way to understand how appointments are being distributed across your team.

A consolidated timeline of the appointment's lifecycle — created, rescheduled, cancelled, completed, or marked no-show — including:

Who or what made the change: a team member, the AI text or voice agent, the calendar widget (the public reschedule and cancel links), or a calendar sync (Google, Outlook, ServiceTitan, HouseCall Pro).

What changed (old → new), with timestamps.

No more digging through separate logs or the contact timeline to reconstruct what happened.

Names are captured at the time of each event, so the history stays readable even if a team member is later removed.

Just open an appointment and view the "History" tab.

We've been working on fine-tuning how Agents access and understand your knowledge base. The new approach is live today. Across testing on data where previous failures to find and focus on the correct answer/information in the knowledge base occurred,

This should result in better use of the knowledge base you have configured!

Remember, if you don't want your agents mentioning something, don't put it in your knowledge base. If you want agents mentioning something without being asked about it, put it in their instructions. and if you want agents to have access to certain info on an as-needed basis, put it in the knowledge base. (FAQs are the best way to do this).

to support richer dynamic variables, including standard contact fields, custom contact fields, account fields, appointment owner fields, and legacy calendar variables.

When editing any of these in your calendar settings tab, simply press the key on your keyboard to see the updated list of available variables.

to user profiles.

Added support in appointment confirmation messages.

Added an explicit setting to include the appointment owner bio image in contact SMS confirmations, sending those confirmations as MMS when an image is attached.

To configure, set the user bio and image in their settings. Doing it for yourself? Go to Settings > Myself. Doing it for others? (if you are an admin) Go to Settings > Team.

Once set you can include the field in the part of the confirmation message settings. tab within a particular calendar's settings. You can check the box below to also include the image.

Voice Agents 3.0 routes every call to the most powerful available model and automatically falls back to close runner-ups if there's an outage — no configuration required.

Best in class transcription which modifies on-the-fly as more of the conversation provides context, yielding the most accurate transcription currently possible.

Watch active voice AI calls as they happen: live transcript, live tool events, live handoffs. When a call needs a human, take it over directly from the monitor. The handoff is smooth — the caller isn't put on hold or rung again — and the human-handled portion is automatically transcribed and appended to the full call record. Just click the button and introduce yourself!

Agents can now look up a calling contact's existing appointments and reschedule them on the spot. Rescheduling keeps the same calendar and team member, validates availability the same way new bookings do, and fires all the standard confirmations and calendar sync events.

Test any agent live in your browser — no phone call required. You'll see the transcript build in real time, watch which tools the agent uses (and what they return), and safely try out booking, field extraction, and conditional tools without real side effects. Plus don't burn credits on testing. Test as much as you need - no credits used when testing in the browser!

Partially a result of better transcription, but also a dedicated model now runs to handle field extraction tools, alleviating your main agent of that work, helping it to stay focused and providing you with more reliable information recording.

Agents can now hand a live call to another AI agent mid-conversation — instantly, without a phone transfer, and with the full conversation history carried over. This replaces the old pattern of phone-transferring between AI agents (which forced callers to re-explain themselves). Use it to route to focused agents: phone-numberless experts that exist purely to handle one area, like billing, scheduling, or a specific product line.

Instead of relying on automated machine detection (which is unreliable across carriers), v3 agents simply listen and recognize a voicemail the way a person would. It's more intuitive and tested better in practice.

Browse and preview voices before choosing one, with photos and short audio samples. Voices are organized into two collections: (high-performing voices from our xAI) and (voices designed by and unique to ). Filter by gender, collection, and language, including multi-language capable voices. No legacy voices will remain on the platform.

Every v3 call log now includes a tab — a chronological timeline of everything that happened on the call with timestamps: transcript turns, tool invocations, interruptions, handoffs, and human takeover moments.

You can migrate your account from the AI Voice Agents page. Once migrated, you'll see updated agent editor options, the new browser test interface, the Transfers & Handoffs tab, and all v3 features immediately. Accounts not yet migrated continue on the legacy platform with no disruption. All accounts will eventually be migrated over (~30 days). Check out the help library articles on voice agents to see more info on new features and when you're ready migrating is just one click away!