Collaborate live on your dashboards
No more screenshotting charts into Slack. Discuss your dashboards where the data lives, in real time, with your whole team. Ask the AI a follow-up and get an answer in the same thread.
Why did this dip on the 14th?
Stripe sync broke that morning. @claribi can you confirm?
Confirmed. Stripe data source returned 0 rows between 8:12–14:30 UTC on May 14. Backfilled at 14:31.
Your team's data discussion is happening in the wrong place
Someone screenshots a chart, pastes it in Slack, and three people reply with questions. By the time anyone answers, the number has changed and nobody knows which snapshot they're talking about.
The discussion should happen on top of the live data, not in a separate tool where every message is already stale.
Screenshot forensics
Three different PNGs pasted into Slack, each showing a slightly different version of the same chart. Good luck reconciling them.
Context lost between tools
The insight is in clariBI. The discussion is in Slack. The decision is in a Notion doc. Nobody can reconstruct why you did what you did last quarter.
Async teams paying a meeting tax
Distributed teams block out a weekly meeting just to "walk through the dashboard." Most of it could be a comment thread.
AI lives in yet another tab
You have to leave the dashboard, open the chat interface, re-paste the context, and hope the AI understands. The question dies of friction.
How live collaboration works in clariBI
Comments anchored to widgets. Mentions that actually work. AI in the same thread.
Comment threads on any widget
Start a conversation directly on a dashboard widget, chart, or report section. Reply, edit, resolve. The discussion stays anchored to the data that sparked it, so six months later you can still find the context.
Threads collapse out of the way when you're focused on the data and expand when you want to read the conversation.
- ✓ Inline reply, edit, and resolve
- ✓ Emoji reactions to save typing
- ✓ Markdown formatting in comment bodies
@-mention teammates
Type @ and pick the colleague you want to pull in. They get a notification immediately — in-app and via email — with a deep link to the exact comment.
Mentions respect workspace permissions. You can only mention people who have access to the dashboard; there's no way to accidentally leak data by pinging an outsider.
- ✓ Autocomplete with avatar and role
- ✓ Email + in-app notification
- ✓ Deep link to the exact thread
@claribi AI mentions
Mention @claribi in any comment to ask the AI a follow-up question. The AI reads the surrounding thread, pulls the live data from the dashboard, and responds directly in the thread.
It's like having a data analyst sitting in every Slack channel — except the analyst actually has the data and never gets tired.
- ✓ Reads the widget's live data, not stale snapshots
- ✓ Understands the surrounding comment context
- ✓ Counts against your AI credit allowance
Real-time WebSocket updates & workspaces
Comments and replies appear instantly. No refresh, no polling. Multiple people on the same dashboard see each other's comments and typing indicators live.
Workspaces group dashboards, reports, and data sources by team. Invite members, manage permissions, share resources — without dumping everything in one giant shared folder.
- ✓ WebSocket push for comments, edits, resolutions
- ✓ Typing indicators like any modern chat tool
- ✓ Workspace-level notification controls
Built for teams that don't sit in the same room
Remote, hybrid, and async workflows.
Distributed teams across timezones
The US team signs off, the EU team signs on. Comment threads let both sides pick up the conversation where the other left it, with full context, no "can you catch me up?" stand-ups.
The dashboard is the handoff document.
Analytics teams supporting stakeholders
Instead of fielding the same questions in five different Slack channels, the analytics team answers once in a comment thread that everyone can see, search, and reference later.
Institutional knowledge builds up on the dashboard itself.
Weekly reporting cadences
The Monday business review happens on the dashboard. Each exec leaves comments on the metrics they own, the AI fills in the "why" questions, and by Tuesday morning the whole week is planned without a single live meeting.
The dashboard is the meeting.
Incident retrospectives
When something goes wrong, the retro happens on the actual data. Comment on the exact spike, @-mention the on-call engineer, ask @claribi what triggered the anomaly. The record lives with the dashboard, not in a lost Google doc.
Post-mortems that anyone can find later.
Built into the rest of clariBI
Comments, mentions, and workspaces share the same permission model, the same AI engine, and the same audit trail as the rest of the platform.
AI-Powered Analytics
The AI that answers @claribi mentions is the same one powering dashboards and insights.
Granular RBAC
Workspace permissions control who can comment, mention, and resolve threads.
Public Sharing
Share the dashboard outside; keep the comment thread private to your workspace.
Real-time collaboration is available on Professional and Enterprise plans.
More collaboration capabilities
The details that make real-time collaboration actually pleasant to use every day.
Emoji reactions
React to a comment with an emoji instead of typing "agreed" for the fifth time this week. Saves keystrokes and signals consensus fast.
Resolve to archive
Mark a thread resolved when the question is answered. It collapses out of the main view but stays searchable for later reference.
Per-thread mute
Mute a chatty thread without unfollowing the whole dashboard. Precision muting — keep the signal, kill the noise.
Deep linking
Every comment has its own shareable URL. Paste it in Slack, in a ticket, or in another comment — clicking opens the dashboard with that thread expanded.
Digest notifications
Prefer one email a day instead of a flood? Digest mode bundles all your mentions and thread activity into a single morning summary.
Presence indicators
See who else is currently viewing the same dashboard. Small avatars in the header make impromptu async conversations feel more like a live room.
Under the hood
Comments and presence events flow over a persistent WebSocket connection. When you post or edit, the update is fanned out to every active viewer of the same dashboard in under a second.
Fallback HTTP polling kicks in if the WebSocket drops, so a flaky network doesn't kill the conversation. Typing indicators use a short-lived ephemeral channel that doesn't hit the database.
The @claribi AI mention is a special comment handler: when the AI is mentioned, a background task pulls the widget's live data, runs it through the AI engine, and posts the response as a reply. The whole round-trip usually finishes in a few seconds.
WebSocket transport
Persistent connection with automatic reconnection. Sub-second delivery for comments, edits, reactions, and presence.
Permission checks
Every incoming WebSocket message is authorized against the user's workspace role. Nothing bypasses RBAC.
AI mention pipeline
Background worker pulls live data, calls the AI engine, and posts the reply. Uses AI credits from your plan's allowance.
Stop screenshotting. Start discussing.
Move the data conversation to where the data lives. Free trial — no credit card.
Getting started with live collaboration
From "pasting screenshots in Slack" to "threaded discussions on live data."
Create a workspace per team
Start by setting up one workspace per team: Marketing, Finance, Product, whatever matches your org chart. Each gets its own dashboards, reports, and conversations without cross-contamination.
Invite members by email; they land directly in the workspace.
Start your first comment thread
Open a dashboard, click any widget, and add a comment. Ask a question, flag an anomaly, or just leave context for next week's review. The thread is anchored to the widget forever.
This is the hardest step — once people do it once, they do it all the time.
Try @claribi in a comment
Ask a follow-up question by mentioning @claribi in the thread. The AI reads the context, pulls the live data, and posts a reply. Great for "why did this drop?" style questions.
Uses AI credits from your plan; watch the usage panel to track consumption.
Move your weekly review into clariBI
The next time you'd normally hold a live meeting to "walk through the numbers," try leaving comments on the dashboard instead. Summon people with @-mentions. Let them respond async. See if you still need the meeting.
Most teams discover they don't.
How we think about collaboration
The principles behind the clariBI collaboration design.
Conversations belong next to the data
The moment a conversation moves from the dashboard to Slack, context starts leaking. Comments live on the widget they're about, so six months later anyone can reconstruct what was discussed and why.
Notifications should be signal, not spam
You get notified when someone mentions you, replies to your thread, or resolves a thread you started. You don't get notified every time someone else has a conversation you weren't part of.
Async first, real-time when it helps
Threads are asynchronous by default — leave a question, come back to the answer. When multiple people are online at the same time, presence and typing indicators turn it into a live conversation without forcing anyone to be there.
The AI is a participant, not a feature
@claribi is mentioned the same way a human is, responds in the same thread, and is subject to the same permissions. Treating it as a participant makes it feel natural instead of bolted-on.
History is not disposable
A resolved thread isn't deleted — it's archived. You can always go back and read the discussion that led to a decision, even a year later. Institutional knowledge compounds instead of evaporating.
Permissions apply to conversations too
Workspace RBAC governs who can comment, mention, and resolve. You can't accidentally pull an outsider into a sensitive thread, and you can scope commenting rights separately from viewing rights.
Collaboration features across plans
What's included at each tier.
| Feature | Starter | Professional | Enterprise |
|---|---|---|---|
| Basic workspaces | Yes | Yes | Yes |
| Comment threads on widgets | No | Yes | Yes |
| @-mention teammates | No | Yes | Yes |
| @claribi AI mentions | No | Yes | Yes |
| WebSocket real-time updates | No | Yes | Yes |
| Presence indicators | No | Yes | Yes |
| Emoji reactions | No | Yes | Yes |
| Advanced notification routing | No | No | Yes |
See pricing for full plan details.
Real-Time Collaboration FAQ
Common questions about comments, mentions, and workspaces.
Does the AI actually understand my data when I @-mention it?
Yes. When you mention @claribi in a comment on a widget, the AI reads both the surrounding comment thread and the live data underlying that widget. It answers with your actual numbers, not a generic response.
Can I export comment threads?
Yes. Comment threads can be exported as part of a report export, or pulled directly via the REST API for archival or compliance purposes.
Are comments private to the workspace?
Yes. Comments are scoped to the workspace that owns the dashboard. Even if you publicly share the dashboard, viewers outside the workspace can't see or add comments.
What if a colleague doesn't have a clariBI seat?
You can't @-mention them directly. For read-only access without a seat, use the public sharing link feature instead — they can view the dashboard but won't participate in comment threads.
How do I turn off notifications?
Per-thread, per-workspace, or globally. Mute a thread when the conversation is done; mute a workspace when you're on vacation. Email and in-app channels can be configured independently.
How many simultaneous viewers can a dashboard handle?
There's no hard limit. In practice, teams of dozens of simultaneous viewers on the same dashboard work smoothly. If you're planning a large all-hands broadcast, let us know in advance so we can keep an eye on capacity.
Does real-time collaboration work on mobile?
Yes. The dashboard viewer is mobile-responsive, and comment threads work on touch devices. Typing a long comment on your phone is less fun than on a laptop, but reading, reacting, and short replies work just fine.
What happens to comments if I delete the dashboard?
Deleting a dashboard also deletes its comment threads. If you need to preserve the discussion, archive the dashboard instead — archived dashboards keep their comments and can be restored.
Can I edit a comment after posting?
Yes. Edits are timestamped and an "edited" indicator shows on the comment. The original content is preserved in the audit log for compliance teams that need the history.