Your AI team in every Slack channel

Installation is a one-click flow from inside your DeskFerry workspace. Go to Settings > Integrations, find the Slack tile, and click Add to Slack. You will be redirected to Slack's standard OAuth screen where you pick the workspace, review the scopes DeskFerry is requesting (sending messages, reading channels you invite it to, posting scheduled reports), and approve the install. Once the bot is added, the DeskFerry app appears in your Slack sidebar under Apps. If you are a Slack workspace admin, you can install it yourself; if you are not, Slack will either route an approval request to your admin automatically, or install immediately, depending on your workspace's app policy. The whole flow takes under a minute, and once DeskFerry is in Slack you can start DMing agents right away — even before you have configured which channels they live in. Uninstalling is equally straightforward: revoke the app from Slack's app directory, or disconnect it from DeskFerry Settings > Integrations, and all bot access is cut off immediately.

Yes — once you invite the DeskFerry bot to a channel, you can @mention any of your agents by name and the bot will respond in-thread. Each agent on your DeskFerry workspace gets its own Slack identity, so @sales-agent, @support-agent, and @ops-agent all behave like distinct teammates with their own tools, data access, and guardrails. When you @mention an agent, it reads the surrounding context in the channel (the last N messages in the thread, plus any file attachments or links) so it can reply with full awareness — no need to re-explain the customer you were just discussing or the bug you were triaging. You can invite the same agent to as many channels as you want, and each channel acts as its own independent context. You can also restrict which agents are @mentionable in which channels from the DeskFerry admin panel — useful for keeping sensitive agents out of broad channels, or giving private channels their own specialized agent.

Only the channels you explicitly invite the DeskFerry bot to, and only when it is addressed. The DeskFerry Slack app follows Slack's standard bot permission model: it does not see any channel it has not been invited to, and in the channels it is invited to, it only reads messages when it is @mentioned or when someone DMs it directly. It does not silently ingest, index, or train on your Slack history. When you @mention an agent, it pulls the surrounding thread context to answer sensibly — that context is used in-session to generate a response and is not persisted beyond standard audit logging. If you want even tighter control, you can scope each agent to specific channels in the DeskFerry admin panel, so it is physically unable to respond outside approved spaces. DMs to an agent are private between the user and the agent and are never visible to other Slack users.

When an agent wants to take a high-stakes action — send an external email, refund a charge, update a customer record, post to a public channel, or anything else you have flagged as needing human-in-the-loop — it posts an approval card to Slack with the full proposed action, the data it will touch, and Approve, Reject, and Edit buttons inline. You click Approve and the action fires in seconds; you click Reject and nothing happens; you click Edit to tweak the payload (say, rewrite the outgoing email or change the refund amount) and resubmit. Every approval is logged with who clicked what and when, so you have a full audit trail. Approval policies are configured per-agent in DeskFerry — you decide which actions require approval, who is allowed to approve them, and how long the agent should wait before timing out. See the human approval docs for the full policy model. Learn more about human approval.

Yes — any scheduled workflow you have set up in DeskFerry can deliver its output to a Slack channel or DM. Common patterns include a daily pipeline update posted to #revenue at 9am, a Friday afternoon support digest to #exec-updates, real-time alerts to #ops when a critical metric breaches threshold, and a morning briefing DM to each individual on the team with their personalized day ahead. Reports are rendered as native Slack messages — not raw JSON or long email walls of text — with headers, bullet points, inline tables where useful, and links back to the full detail inside DeskFerry for anyone who wants to dig in. You can stack multiple schedules per agent, route them to different channels, and change cadence any time without touching code. See the schedule work docs for the full scheduling model. Learn more about schedule work.

The DeskFerry Slack app works on every Slack plan — Free, Pro, Business+, and Enterprise Grid — because it uses standard Slack bot APIs rather than any paid-tier-only features. On Free Slack workspaces, you do hit Slack's own 90-day message history limit, which can affect how much channel context an @-mentioned agent can pull from; Pro and above give you full history. On Enterprise Grid, the DeskFerry app can be installed once at the org level and managed across all workspaces from a single DeskFerry admin panel, with per-workspace channel scopes if you need them. SSO, SCIM, and audit log export all work as expected on Enterprise Grid. If your team is on a custom Slack contract with unusual restrictions (restricted app installs, data residency requirements, custom DLP tooling), reach out to us and we will work through the install with your admin directly.