Brand new 'Reservations' menu item (available both on customer side and agent side):

Brand new 'x10 AI' tab with two menu items:

Configuration menu item:

Knowledge Enrichment menu item:

Once customers answer enough questions, they can run the assessment again on the Configuration menu item and get updated proactive recommendations on automation modes.

We expect AI automation coverage to significantly increase for the select customers using this feature

Agents had to manually write and send guest inquiries and operational tasks outside of platform, which is a time consuming process that also isn't streamlined in terms of tracking.

For select customers, our agents are now able to create tasks (both guest inquiries and operational tasks) directly from a conversation.

For operational tasks (screenshot below), those can be synced to Breezeway (and receive updates from Breezeway).

For guest inquiries (screenshot below), those can be synced to Slack.

Tasks can be tracked via the conversation itself

The new Tasks section in our navigation menu also allows users to not only track tasks but also conduct advanced actions such as filtering, search, copying the task, and updating/receiving updates depending on the type of the task.

This isn't a fresh coat of paint. It's Extenteam shedding its skin it outgrew for one that matches where the product is actually going: more modern, more AI-forward, built to hold everything we're shipping next.

On the customer side, every section has been redesigned. To name a few:

It was highly requested by agents that "send on enter" is removed, so this was done. At the moment if enter is clicked on the keybload the message continues in a new line. Sending a reply can be done either on ctrl/cmd + enter or just by selecting the "send" button in our UI.

Same features, same data, same functionalities, just easier to find, read, and move through. Nothing you've set up changes.

Cleaner type, consistent spacing, and a visual language that finally matches the quality of the product underneath it, so the interface gets out of your way instead of asking users to decode it.

When x10 detected an escalation or opened a task, it went quiet. It would leave an internal note for the agent but send nothing to the guest. This has been deliberate, since those conversations are sensitive, but it meant the guest who just reported a leak or a broken lock got silence until an agent picked it up.

Acknowledges the guest, then steps back. Guest gets a reply confirming their issue is being looked into. x10 pauses and the agent takes the lead.

This is grounded in production data. Prompts built from real guest messages and agent replies, then scored against human-approved expected output.

Silence after a guest reports a problem is the worst possible response. Now they know they've been heard while an agent works it and a meaningful share of internal-note-only conversations become real guest replies, which we expect to move the automation rate up.

So far, once x10 was paused for a conversation, either by an agent stepping in or by x10 pausing itself, it almost never comes back on. Someone would have to manually re-enable it, and in practice that rarely happens. So conversations that x10 could be handling just sit paused.

◦ Automation is enabled at the property and company level.

◦ There's no pending escalation.

◦ There's no open task tied to the conversation.

Paused conversations were quietly dragging down our automation rate, not because x10 couldn't help, but because nobody flipped it back on. Now x10 self-corrects: it picks conversations back up the moment it's safe to, so guests get answered faster and the onus is no longer on the agents to turn x10 on/off on a conversation. We expect this will result in a higher automation rate, more automated responses and decrease the load on agents

SMS was the one channel x10 AI couldn't fully cover. When a guest texted, there was no reservation attached to the conversation: no property, no dates, no context. We now match reservation details to SMS numbers for all companies, and x10 can automate those texts for select companies once a match is made.

This expands x10's AI coverage to SMS for select companies when we are able to match a reservation. This means higher AI coverage and a bigger share of inbound questions get answered automatically instead of queuing for an agent. More channels under autopilot = higher overall automation rate, and a clearer path to pushing AI coverage up across the board.

Huge shoutout to Nicolás, Giovanni and the team!

Every property management company organizes their custom property info differently and x10 AI wasn't able to read all of them. On top of that, some companies' most important fields crams gate codes, parking, bed setups, and urgent arrival notes all into a single block of free text. So x10 AI used to skip it entirely rather than risk a wrong answer. Now it can read it cleanly. 7,690 properties just got a lot smarter.

Across select companies, x10 AI answers questions it used to bounce to an agent:

Property updates flow through automatically. No agent action needed.

x10 AI now handles more questions on its own. Faster guest replies, fewer escalations, less repetitive lookup work for agents. starting now.

We also built a reusable way to teach x10 how to read any customer's custom fields. Next time a customer stores info in their own custom format, we add a template instead of writing new code.

Huge shoutout to Symeon Esdras and the team!

One of the largest single feature releases in the history of our product is now live. The foundation of our platform is the knowledge that we have about properties and the operations of our customer's companies. In the past months, we released quite a few features for property knowledge. Workflows is all about the operations and automating certain scenarios and communication along the way.

Before, our platform had "Actions". Actions existed for our customers to decide what our team could handle on their behalf and how to handle them. Here are some screenshots of what these looked like.

The actions table.

What it looks like for a customer to configure an action.

How actions appear in the inbox.

Workflows isn't a UI/UX revamp of Actions. This is completely reimagined and rebuilt from the ground up. Here is what is different:

How users can view all their Workflows in our platform

How users can configure their Workflows

How Workflows appear in the Inbox and the UX that guides them

Cheers to the entire team for their efforts in delivering this massive improvement to the platform!

Earlier today we launched some (long requested) features into the inbox product! These features are really focused on bringing immediate efficiency to the multi-tasking nature that is managing the communication operations for short term rental operators. Also, is is very exciting because these features will act as a foundation for other exciting future releases :)

As we all know, guest communication doesn't start and stop with the guest. There are side quests to gather more information, get approvals, etc. Linked Conversations brings that directly into the view of the guest conversation for easy categorization and multi-tasking management. Can't wait to see all of the benefits this brings!

If you watched the demo video above, you already saw how this works. If not, here is the TLDR. We now allow for users to place conversations On Hold as they are waiting for a different action to take place. Then, we listen for any message from the guest OR from Linked Conversations. If something comes in, we move that conversation back to Active so that it can be handled in a timely fashion.