The teammate who’s read everything

An agent inside your workspace that knows your product, your customers, and your numbers. Ask for a metric, a brief, or a reply — it’s already halfway done.

One brain. Every job.

Copilot sits inside your inbox and already knows the answer — the metric, the queue, the language, the reply.

Ask anything

Which channel spiked? Ask in plain words — the answer comes back with the numbers, live from your workspace.

Knows what’s missing

Copilot mines your conversations for questions no article answers — counted, ranked, and drafted for your knowledge base.

Already read

Open a 50-message thread and the brief is waiting — and it rewrites itself on every new message.

Both sides native

Messages arrive in your language, replies leave in theirs — nobody notices the border.

Drafts the reply

Copilot writes the answer from your own sources — you read it, tweak it, send it.

One inbox, one knowledge base, one quota

The same sources power Copilot drafts and your customer-facing AI Agent. Copilot actions draw from one shared AI-requests pool — no per-seat add-on.

Give your team a copilot — free for 14 days.

Start 14-day trial

Before you turn it on

What Copilot does today, how it's billed, and where early access is headed.

Today Copilot automatically summarizes every conversation into a 3-bullet brief, translates inbound threads into your operator's language with a one-click "Show original" toggle, and translates your outgoing drafts into the customer's language. AI reply drafting and free-form knowledge-base questions are in early access — book a demo to join.
Every time a customer sends a message, Copilot regenerates a background summary of the last 50 messages: what the customer needs, what was offered, and where things stand — three bullets, each 15 words or fewer. Summaries appear above the thread and in the case details, written in each operator's profile language. Bot-handled threads and Zendesk-synced tickets are covered too.
Operators choose from 15 interface languages for summaries and translation. On the customer side, Copilot detects the language automatically and translates in both directions — inbound threads into your language, outbound drafts into theirs. Up to 50 messages are translated per pass, and translations are cached so threads reopen instantly.
Yes — that's the core of the early-access drafting feature. Reply drafts are generated with retrieval over the knowledge sources linked to your agent, and every draft cites the sources it used so operators can verify before sending. The same knowledge base powers your customer-facing AI Agent, so there's nothing extra to maintain.
Copilot actions draw from the same unified AI-requests quota as your customer-facing AI replies, one for one. There's no separate per-seat Copilot add-on and no hidden meter — one pool, visible in your usage dashboard.
Yes. Copilot runs on the conversation itself, not the channel: summaries auto-generate for messages arriving from your chat widget, email, and Zendesk Support sync alike, and translation works on any thread in the inbox.

Stop opening threads blind

Summaries and translation are live. Start a free trial and open your first thread briefed.