Dotfile <> Slack Integration
Last updated: September 16, 2026
A guide to receiving a Slack message at every stage of an account opening from draft through to closed plus a message whenever an AML hit needs review.
1. How Integration Studio works
Integration Studio is a workflow builder embedded directly in Dotfile. You build recipes: each recipe has one trigger (an event in Dotfile) and one or more actions (what happens as a result here, posting to Slack).
The Dotfile side of the integration is already connected and scoped to your workspace. The only connection you need to create yourself is Slack. More information on Integration studio can be found here.
2. Before you start
Admin access to your Dotfile workspace.
Permission to install a Slack app in your Slack workspace. If your Slack is locked down, your Slack admin will need to approve the connection once worth lining up in advance, as it is the most common thing that blocks this setup.
The Slack channel you want to post into. A dedicated channel such as #kyb-notifications is easier to live with than a busy general channel, because these messages can be frequent.
3. Connect Slack
Open Integration Studio from your Dotfile workspace settings.
Go to Connections and create a new connection, selecting Slack.
Authorise the Slack OAuth prompt, choosing the Slack workspace to post into.
Invite the Slack app to your target channel in Slack, run /invite @<app name> in #kyb-notifications. Posting to a channel the app has not joined is the second most common failure, and it fails silently in some configurations.
Leave the existing Dotfile connection alone. It is already authenticated.
4. Recipe 1 — notify on every status change
This covers open, approved, rejected and closed.
In the Integrations project, create a new recipe.
For the trigger, pick the Dotfile connection and select the event Case.StatusUpdated.
Add a Slack action: Post message to channel. Choose your channel and compose the message from the trigger's output fields, for example:
📁 Case {{case.name}} is now {{case.status}}
Previous status: {{case.previous_status}}
Open in Dotfile: {{case.url}}
Save, then Start the recipe.
Field names in the message body must match what the trigger actually outputs build them using the field picker rather than typing them by hand, since the picker guarantees the correct path.
Case.StatusUpdated fires on transitions. A case is created already in draft, so entering draft is not a transition and this recipe will not fire for it. To be notified at the very start of an opening, use a second recipe triggered on Case.Created (section 5).
This is the one part of the request that needs two recipes rather than one, and it is easy to miss until you notice the draft notifications never arrive.
Filtering to specific statuses
If you only want some statuses, add a condition step after the trigger — for example continue only if status is approved or rejected. Alternatively, use a switch to send differently worded messages per status: a rejection probably deserves a different tone and a different channel from an approval.
Related events you may also want
Event | Fires when |
Check.Rejected | A check is rejected, automatically or by a reviewer |
Check.Approved | A check is approved |
Check.Expired | A check expires — useful for documents with an expiry date |
Check.Started | A check begins running |
Case.RiskUpdated | The case risk level or score changes |
Case.FlagsUpdated | A case flag is added or removed |
Case.ReviewDue | A periodic review falls due |
Case.RiskUpdated | is worth considering alongside the hit notification: a hit that pushes a case to high risk is usually the thing a compliance team wants to see immediately, and the risk event carries that context where the check event does not |
5. Full event reference
Events are organised as parents with children. Subscribing to a parent gives you all its children; subscribing to a child narrows it down.
Case - Case.Created, Case.Deleted, Case.ReviewConfirmed, Case.ReviewDue, Case.ReviewUpdated, Case.ClientPortalWaitStepTriggered, and Case.Updated with children Case.StatusUpdated, Case.TagsUpdated, Case.FlagsUpdated, Case.RiskUpdated, Case.InfoUpdated, Case.TemplateUpdated, Case.MetadataUpdated, Case.AssigneeUpdated, Case.ContactHasActionsUpdated, Case.ReviewerHasActionsUpdated
Check — Check.Started, Check.ReviewNeeded, Check.Approved, Check.Rejected, Check.Expired, Check.Deleted
Company — Company.Created, Company.Deleted, Company.Updated (children: Company.InfoUpdated, Company.MarkedAsRelevant, Company.MarkedAsNotRelevant)
Individual — Individual.Created, Individual.Deleted, Individual.Updated (children: Individual.InfoUpdated, Individual.MarkedAsRelevant, Individual.MarkedAsNotRelevant)
Other — CaseReport.Generated, DocumentOrder.Completed, DocumentOrder.Failed, Note.Created, Note.Updated, Note.Deleted, NoteComment.Created, NoteComment.Updated, NoteComment.Deleted, AutonomyChatRun.Completed, AutonomyChatRun.Failed
Subscribing to the Case.Updated parent will produce a lot of Slack traffic, since it includes metadata and info changes. For notifications, prefer the specific children.
6. Testing
Start with sandbox.
Create a test case and walk it through the statuses you care about.
Check each recipe's job history in Integration Studio. Every trigger produces a job, and a failed job shows the exact error — this is the first place to look when a message does not arrive.
Common causes of silence: the recipe was saved but never started; the Slack app is not a member of the channel; a condition step is stricter than intended.
7. Practical limits
Volume. Status changes on a busy pipeline generate a lot of messages. A dedicated channel and specific child events rather than parents will keep this usable.
Plan. Integration Studio workspaces start on a default plan. If you need a large number of recipes or higher task volumes, that may need to be raised commercially rather than configured.
Per environment. A second Dotfile environment means a second Integration Studio setup, its own project and its own recipes. Nothing is shared.
Ordering. Events are independent. Two status changes in quick succession may not arrive in Slack in the order they happened.