Front Collision Detection & Presence Explained (2026)

What collision detection actually does

Collision detection is the real-time signal that someone else is already on the conversation you just opened. Per Front's guide to real-time collision detection, Front automatically notifies teammates when another person is working on a message in a shared inbox — and specifically when they are replying to it. You don't have to enable anything or ask; the moment a colleague starts drafting, the interface tells you.

The point is coordination without conversation. Nobody has to Slack "hey, are you taking the Acme ticket?" because the inbox already answers that question visually. The two failure modes a shared inbox invites — two people answering the same email, or nobody answering it because everyone assumed someone else would — are exactly what these signals defuse.

Two things make the signal trustworthy rather than merely decorative. First, it's real-time: presence updates as people open, read, and start typing, not on a refresh. Second, per the same Front documentation, draft content is automatically shared and updated immediately — so when a teammate is mid-reply, you're not just told that they're replying, you can see what they're writing. That shared-draft surface is where a live collision announces itself.

The screenshot above is the composer where a collision surfaces — the Shared draft — Editable by teammates indicator. On a single-seat account no live "another teammate is viewing" badge can appear, because collision detection is a collaborative feature: it needs at least two teammates in the same inbox before there's anything to collide. What you're seeing is the honest presence surface, with the collision badge described from Front's docs below.

The participants menu: who has seen it

Beyond the live "someone is typing" alert, Front keeps a running record of who has touched a conversation. The participants menu lists the avatars of every teammate who has participated — anyone who has commented, @mentioned someone, replied directly, or subscribed to the conversation.

The avatars carry read-state, which is the useful part:

1. Color avatars mean that teammate is reading the conversation right now, or has already read the latest message or comment. They're caught up.
2. Black-and-white avatars mean there's new activity in the conversation that person hasn't seen yet — Front marks them Unread.

So at a glance you can tell not just who cares about a conversation but who is current on it. If you're about to hand something off, the participants menu tells you whether the person you're handing it to has actually seen the latest reply. It's presence for the whole lifecycle of a thread, not only the live moment.

Teammate status colors: who's even around

Collision detection tells you who's on this conversation. Teammate status tells you who's available at all — and Front wires the two ideas together so routing respects reality. Per Front's teammate status colors reference, each teammate carries a colored presence dot:

This matters for collision avoidance in a subtle way: status doesn't just decorate an avatar, it changes routing. A teammate marked Busy stops receiving new rule-based assignments, and one marked Out of office gets conversations unassigned when new activity arrives. That's the flip side of collision detection — instead of two people landing on one conversation, status ensures work isn't handed to someone who has stepped away. Front surfaces these colors wherever you'd act on them: when assigning a conversation, when @mentioning a teammate in a comment, in the live analytics dashboard, and in the sidebar teammates list.

How the pieces work together

None of these surfaces is meant to be used alone; the collaboration model is the sum of them. In practice, a well-run Front inbox leans on three coordinated signals.

Assignment is the first line of defense: when a conversation is assigned to one teammate, it's clear who owns the reply, and status colors keep rules from assigning to someone who's away. Collision detection is the safety net for everything that isn't formally assigned — the moment two people happen to open the same unassigned conversation, presence and the shared draft make the overlap visible before either sends. The participants menu is the audit trail, showing after the fact who saw what and who's caught up.

Together they're why a Front team can share one inbox without a spreadsheet of who's doing what. This is the shared inbox model working as designed, and it's the same collaborative fabric behind shared drafts and comments and @mentions — internal coordination that the customer never sees. For teams splitting one platform across support, sales, and ops, it scales through team inboxes, each with its own presence.

The honest limits — and where an AI layer picks up

Collision detection is genuinely one of Front's best features, and it's worth being clear about that before naming its edges. It's built into the shared inbox, it's real-time, and it does exactly the coordination job it claims to.

But its whole nature is also its ceiling. It's a collaborative feature by definition — it needs 2+ teammates to do anything. On a solo account or a single-seat trial, there's nobody to collide with, so the live badge simply never appears (which is why the screenshot above is honestly the presence surface, not a live collision). More fundamentally, collision detection solves a human coordination problem: it stops two people from answering the same email. It does nothing about the underlying cost — that a person has to answer the email in the first place. Presence tells you who is replying; it can't read the customer's question, look up their order, and write the reply for you. And the coordination signals themselves lean on your busier collaboration and permission features, some of which are plan-gated (granular workspace permissions and SSO sit on higher tiers). Front's own pricing tiers lay out where each unlocks.

That's the seam an AI agent layer fits into — not to replace collision detection, but to reduce how often two humans need coordinating over the same routine reply at all. The broader category of AI agents for customer service exists for exactly the work presence can't do: understanding the question and answering it. Macha is one such layer. It runs on top of the Front you already use through the live Macha–Front connector — it does not replace Front, your shared inboxes, your assignments, or collision detection. Your team keeps its presence signals and its safety net. Macha's agent reads an incoming conversation, understands intent rather than status, and drafts or sends a grounded reply — pulling a real order or account detail through a custom tool that turns your REST API into something the agent can call. Collision detection means two teammates don't both answer a shipping question; an agent means, for the routine ones, neither of them has to. And Macha's credits are consumed per AI action, never per resolution — reasoning and coordination are different costs, and it's honest to price them apart.

The clean division of labour: let Front's presence keep your humans from tripping over each other, and layer an agent on top so fewer conversations need a human at all.

FAQ

How does Front collision detection work? Front automatically notifies teammates in real time when someone else is working on — or replying to — a message in a shared inbox. Draft content is shared and updated immediately, so you can see that a colleague is mid-reply (and what they're writing) before you send your own, preventing duplicate responses.

Do I need multiple teammates for collision detection? Yes. Collision detection is a collaborative, shared-inbox feature — it only has something to show when two or more teammates are in the same inbox. On a single-seat account there's no one to collide with, so no live "another teammate is viewing" badge appears; you'll still see the underlying presence and shared-draft surface.

What do the teammate status colors mean in Front? Green means Available (can receive new rule assignments), Yellow means Busy (pauses new rule assignments), Red means Out of office (mutes notifications and unassigns conversations on new activity), and Grey means Disconnected (no device connected to Front's push system). Status appears when assigning, @mentioning, in the analytics dashboard, and in the sidebar teammates list.

What does the participants menu show? It lists avatars of every teammate who has viewed, commented, @mentioned, replied to, or subscribed to a conversation. Color avatars mean they're reading it now or have read the latest message; black-and-white avatars mean there's unseen new activity marked Unread.

Can I add AI to Front without replacing collision detection? Yes. An AI agent layer like Macha connects to Front as a live connector and runs on top of your existing shared inboxes — collision detection, presence, and assignments all keep working. The agent handles the reasoning-heavy part (reading and answering routine conversations) so fewer of them ever need a human to coordinate over.

Ready to have fewer conversations that need a human at all? Start a free trial of Macha and connect it to your Front in minutes.