The morning shift is starting. Who is arriving today? Which guests still need a room? Are the assigned rooms already clean? Answering these questions often means opening several screens and putting the information together yourself.
In Discontinue, you start with a message in chat. Ask to see the data, follow up on the result and handle the next steps in the same conversation. Once the process works for you, turn it into a report or an agent for recurring work.
We will follow one example: preparing today's arrivals, then creating a daily arrivals overview and an agent to handle room assignments.
1. Connect the systems you want to work with
Open your workspace and connect the systems you need under Connectors. Our example uses Apaleo. If you also want to exchange data with Google Sheets, connect Google Workspace too.
The tasks Discontinue can perform depend on the connected systems and the permissions you grant.
If results should later reach your team by email, Slack or Teams, set up the destination under Channels. Send a test message to check that everything arrives in the right place.
2. Start with the task you need to solve today
Describe what you need in chat. You do not need to plan a report or an agent yet:
Show today's arrivals for my property as a table here in chat. Ask which property if more than one could apply. Include the reservation reference, arrival time, booked room category and assigned room. Put arrivals without a room first.
The result appears in the conversation, ready for you to build on:
Add the current cleaning status for assigned rooms. Separate guests who have already checked in from those we are still expecting. Show “unknown” wherever information is missing.
Step by step, you shape the overview your team actually needs. Add columns, ask about a particular reservation or change the order. Compare a few entries with Apaleo to check that the details are correct.
3. Handle the next steps in the same conversation
The overview shows what still needs doing. You can ask Discontinue to act on it directly. Select a specific reservation from the table:
Find a clean room in the booked category that is available for this reservation's entire stay. Show me your proposal and assign the room after I confirm. Keep the rate and stay dates unchanged.
Discontinue carries out the action using the permissions you have granted. Afterwards, ask to see the updated overview and check the assignment in Apaleo.
You can also use chat to transfer data between connected source and destination systems. If your team keeps a handover list in Google Sheets, for example, you might write:
Add the outstanding room assignments to our connected Google Sheets handover spreadsheet. Ask me which spreadsheet and sheet to use. Show me the rows you intend to add before writing them.
You have now completed a real task. The next step is to decide which parts should run regularly.
4. Decide what you want to reuse
A report delivers a recurring analysis. Think of the arrivals overview, ready every morning with current data. AI helps build the queries, calculations and presentation. Each subsequent run follows the saved definition without another AI run. Reports read and process data; they cannot change reservations or perform other actions in your source systems.
An agent handles a recurring task. For example, finding a suitable room for each new reservation and assigning it within your rules and permissions. It uses AI on every run to work through the task in the current situation. An agent can also simply check something and notify the team when it needs attention.
An agent can also handle transfers between systems: for example, reading data from Apaleo and adding selected entries to a connected Excel or Google Sheets spreadsheet.
Our example gives us two jobs: the report provides the morning shift with an overview, while the agent handles room assignment.
5. Turn your overview into a report
You do not need to explain the task from scratch. Continue from the result you have already worked through in chat:
Create a report from this arrivals overview for the selected property. On each run, show arrivals for the current local day. Keep the columns, the split by arrival status and the flags for missing room assignments. Include the current cleaning status, data source and generation time.
The preview appears alongside the conversation. Review it and describe any changes directly in chat:
Show the arrivals that still need our attention first. Put the complete list underneath.
Once the overview works for you, save the report in Discontinue. You can open it there at any time and have it delivered as Excel, PDF or CSV when needed. Choose the recipients, format, channel and time zone, then set up daily delivery at 06:30. Check a test delivery and activate the schedule when everything is ready.
Your team now gets the same clearly structured overview each morning, using current data.
6. Turn the recurring task into an agent
You can build the agent from the same conversation:
Create an agent to handle room assignments for the selected property. When a new reservation arrives, check whether it already has a room assigned. If it does not, find a room in the booked category that is available for the entire stay. Use the allocation rules we discussed and ask about anything still missing.
Show me the proposal for approval before assigning the room. If no suitable room is available, notify the front office with the reservation reference and the reason. Keep rates, stay dates and reservation status unchanged.
Open the created agent from the chat menu in split view. The conversation and the agent stay visible side by side. You can edit it directly, activate the reviewed version and run it manually there.
Check its permissions too. For this example, a workspace admin sets room assignment to Ask first, so the agent requests your approval before making the change.
Review the test run and refine the agent in the same chat:
If several rooms are suitable, prefer one on a lower floor. Keep existing room assignments.
Apply the change and test again. Also check how the agent handles a reservation that already has a room or one for which no suitable room is available.
Automatic triggers require separate activation. Building an agent does not start it running automatically. Once the task, permissions and test result are right, activate the trigger for new reservations.
7. Keep improving it as you work
The first deliveries and runs show whether the result works for your team. Perhaps the overview needs another column. Perhaps the agent needs an additional rule.
Open the report or agent alongside chat again and describe the change. You can inspect the result, edit it and run it again directly. That lets you improve the process as part of everyday work.
Start with a task that needs doing today. Solve it in chat. Turn it into the report or agent that takes work off your team's hands tomorrow.
Get started with Discontinue · Discontinue manual work.