Agents
An agent is an AI worker with standing instructions. It starts on a schedule, on an event from one of your systems, or when you run it yourself. It works through its instructions against your live data, decides what matters, and notifies your team through your channels.
Where chat answers one question and a report delivers the same structure on a schedule, an agent evaluates the current data on each run and decides whether anything needs to be done.
How an agent works
An agent takes over recurring work too variable for a fixed rule. A run follows the agent's reviewed instructions and can execute its saved scripts, query connected systems, and decide what to look up next. This lets you ask for things a fixed rule cannot express: "flag anything unusual", "tell me what changed and why it matters".
It acts only within the permissions you grant, and every run is bounded and recorded step by step.
Creating an agent
A draft is not an activated agent. The builder can rehearse it while editing; automatic triggers stay paused until you start them.
Describe it in chat
New agent on the Agents page opens a conversation with the agent builder, which drafts the whole agent from your description: trigger, applications, instructions, and notification rules.
Review the draft
The draft appears in a panel next to the conversation, and every change you ask for updates it there. Unactivated drafts do not occupy a plan slot. Builder rehearsals are separate from running an activated agent.
Triggers
An agent runs when one of its triggers fires:
| Trigger | When it runs |
|---|---|
| Manual | You press Run agent, with optional one-off instructions and files |
| Schedule | Daily, weekdays, specific days, or monthly, at fixed times or intervals, in the workspace time zone |
| System event | A matching event arrives from a connected system, such as a new Apaleo reservation |
| Webhook | Any external system calls the agent's signed URL |
Webhook setup is copy-ready on the agent's page: the URL, a signed curl example, and a Postman collection. Webhook runs are capped per agent (240 per hour by default); calls over the cap are refused with a signal telling the sender when to retry.
Instructions, tools, and notifications
The Agent tab shows the agent as four questions: when does it activate, what can it access, what does it do, and who hears from it. Everything is editable in place.
- Instructions are plain language: what to examine, how to judge it, what to produce. Workspace-wide AI instructions apply on top.
- Applications: only the connections you attach, each summarized at a glance as read-only, acts with your approval, or acts on its own.
- Tools: a Python workspace for real computation and file work. It never holds your credentials.
- Skills: which documents the agent can produce: Excel, PDF, Word, PowerPoint.
- Reports as data sources: deterministic, structured numbers to reason over, the same figures your team sees. The agent runs the reports granted to it on the Discontinue integration (below).
- Who gets alerts about this agent: the agent's Notify list, the members and channels that are asked about its approval requests and told when a run fails or, for the entries that choose it, after every run. It starts with the agent's creator; anyone in the workspace can change it under Edit, and changing it needs no new version.
- Delivery: the agent sends only to the channels granted to it on the Discontinue integration (below), of any type; a webhook channel receives its message as JSON or as a Teams card, without files. When it sends is part of its instructions, such as "only if something needs attention". One failing channel never blocks the others.
Agents and channels
An agent delivers its results through the Discontinue integration: Discontinue itself, offered to agents like any other integration. Every workspace has it; it needs no connecting and is not listed under Connectors. It acts in your other systems through their own integrations.
Its first area is your channels. An admin grants an agent the channels it may use, in the agent creator or on the agent's page, and the agent sends there only: it lists the channels it was given, and sends to each with the operation of the channel's type. An email is written by the agent (a subject and a formatted body, with attachments) and sent in your workspace's branded email; Slack, Telegram and FTP get the message or file as written; a webhook gets the message as JSON or as a Teams card. A send is an action like any other: it asks first unless you allow it, it shows on the run as a delivery, and it counts against the monthly delivery limit.
| Area | Operation | What it does | Kind |
|---|---|---|---|
| Channels | channels.list | List the channels | Read |
| Channels | channels.send_email | Send an email | Action (asks first by default) |
| Channels | channels.send_slack | Post to Slack | Action (asks first by default) |
| Channels | channels.send_webhook | Post to the webhook | Action (asks first by default) |
| Channels | channels.send_ftp | Upload a file to FTP | Action (asks first by default) |
| Channels | channels.send_telegram | Send to Telegram | Action (asks first by default) |
Renaming a channel renames it on every agent that may use it; a deleted channel stays on the agent's grant, marked as deleted, and sends nothing.
Agents and reports
The second area of the Discontinue integration is your reports. An admin grants an agent the reports it may run, in the agent creator or on the agent's page, and the agent runs those only: it lists the reports it was given, runs one when it needs the figures as of now, and reads the result page by page. A run is an action like any other: it asks first unless you allow it, and it shows among the report's runs and under Report Runs on the agent's page. Reading a run the agent started costs nothing and asks nothing.
| Area | Operation | What it does | Kind |
|---|---|---|---|
| Reports | reports.list | List the reports | Read |
| Reports | reports.run | Run a report | Action (asks first by default) |
| Reports | reports.read | Read a report run | Read |
Renaming a report renames it on every agent that may run it; a deleted report stays on the agent's grant, marked as deleted, and cannot run.
You edit an agent's channels and reports where you edit any integration's access: open the Discontinue integration's tile under Applications on the Agent tab and pick them there. Agents that had channels or reports before they moved into the Discontinue integration were moved over automatically, with the same channels and reports, allowed to send and run without asking as before. Until that has happened for an agent, its page lists its channels and reports read-only.
Permissions and approvals
Safety has two layers: access (the agent never holds your credentials, only a short-lived token limited to the scopes its applications were granted) and permissions (what it may do with that access). Activating a version never changes permissions by itself. Open an application's tile on the Agent tab to set each action to one of three levels:
- Allow: the agent acts on its own.
- Ask first: the run pauses for approval.
- Block: the agent cannot use it.
A run that hits Ask first parks as Waiting for approval. The request shows on the run page and in the Activity feed, and the agent's Notify list hears about it: people by email, channels as a short message, each with a link to the request. Decisions are taken in Discontinue only, on the run page, in Activity, or on the card in chat. Allow it for this run, allow it permanently, or deny it with a message the agent reads.
- Action approvals, an agent asking to use something it has been given, are decided by any member of the workspace: allow for this run, or deny.
- Permission requests, an agent asking for access its connection does not have yet, and Always allow on any request are decided by workspace admins and owners.
A request nobody decides within 24 hours expires and is treated as denied for that run; it is not remembered as a block. Only workspace admins can change permissions.
Runs and history
Every run is recorded step by step. The Runs tab lists them all. Opening one reads like the exchange it was: the email or instruction that started it, each message the agent sent back, the approvals people decided, and who on the Notify list was told; the steps in between sit under Show steps. Earlier and later emails of the same thread link to their runs. A running agent can be stopped mid-run.
-
Outcome: after successful execution, supported runtimes can ask the agent to state how it went, with a two-sentence summary and up to three issues, each with a cause such as an expired credential or missing data.
- Completed the run did what the instructions asked.
- Warning the work was done, but something needs a look.
- Failed the run did not deliver, for example because a system it needed was unreachable.
The Runs tab shows the distribution across recent runs. This closing step is billed like any other turn and can be switched off per agent on its Limits tab. A successful run may show No outcome if the step is disabled, unsupported or produces no valid statement. Failed runs can have a system-recorded outcome.
-
Failures are explained in plain language, with what to do next. Actions completed before a stop are not rolled back; the trail shows what changed.
-
The Audit tab lists every call to your systems: what was read, changed, or prevented. The workspace-wide Activity feed shows everything across agents and reports.
Versions
Each proposed version is kept as a numbered, immutable snapshot, and the active one is the behavior the agent runs. Open any version to see the agent exactly as it defines it, differences highlighted, and activate it to roll forward or back. Older versions are kept, not overwritten.
Edit settings directly on the Agent tab, or use Edit with AI: the builder proposes a new version for you to review and activate. Duplicate agent copies the active version into a new draft; deleting an agent removes it with its run history.
Budgets and cost
Agents are bounded at two levels:
- Per run: a tool-call budget (60 calls by default) and a time budget (10 minutes by default) limit how much a single run can do.
- Per month: every agent has its own euro spend cap (10 € by default) on its Limits tab, alongside the workspace-wide limit. An agent that reaches its cap is blocked; other agents are unaffected.
Agent runs use AI credit, and so does the builder's work when creating and editing agents. See Limits & cost controls for how caps behave when reached.