Skip to content

Labor · Operations

Operational notifications that lead to the right action

We prepare technical states so that the people responsible can see what is new, what can wait, and where action is needed. Follow-up questions and personal settings flow back through the same channel.

Practical setup 01Tested in our own operations

Classifying disruptions clearly and reporting them with purpose

In unattended processes, stalled operations often only come to light once someone notices the missing result. Technical error codes rarely help the people responsible.

Test bench

10 events classified, 2 of them sent to the phone.

Controls

Switch and compare what reaches the people responsible.

Sent to the phone

Everything stays in the log. The level only decides what actively reaches someone.

Event

Click a message on the phone or an entry in the log.

Events

10

all in the log

Sent to the phone

2

2 of 10

Of these, need action

2

the rest needs nobody

Example events with well-known protocol codes, no operational data. The flow abstracts a notification route from our own operations.

How the setup works Play ten example events. Each one lands in the log, errors and notices become plain-language messages. Switch between error code and plain language and choose what may reach the phone.
Scope and limitations

Not part of this demo: This shows the notification path, not frequency, response time, or the operational distribution of errors.

The setup abstracts a process from our own operations and contains no operational or customer data.

Practical setup 02Tested in our own operations

Spotting changes in the daily status picture

Individual event notifications do not show whether a condition is new, whether it is growing, or whether the same known anomaly has persisted for days.

Test bench

5 reports in the inbox. Click a day.

Signals per day

Delivery delayedInvoice rejectedSlow responsesArchive almost full

Click a cell to change its level. The subject lines recalculate right away.

Inbox

Report

Click a day or an email. The comparison with the previous day appears here.

Days “all quiet”

1

of 5 days

Days with a new error class

2

worth a look

Growing issues

1

compared with the previous day

Example week without operational data. The levels show a pattern, not measured quantities. Click a cell to change the week yourself.

How the setup works Play an example week. Each report's subject comes from the comparison with the previous day. Click cells to change the week and see when a new error class stands out.
Scope and limitations

Not part of this demo: Operational figures, recipients, and run intervals stay outside the picture.

The setup abstracts our own status picture without exposing internal metrics or infrastructure details.

Practical setup 03Tested in our own operations

Steering notifications directly by replying

Notification settings usually live in separate interfaces. Follow-up questions and personal delivery preferences end up bypassing the actual conversation.

Test bench

Conversation

  1. Status Tue: all quiet. Just reply to this message if you want to change something or ask a question.

Try these

Settings

Frequency
a report every weekday
Away until
not set

The next seven days

  1. MonReport
  2. TueReport
  3. WedMessage with action
  4. ThuReport
  5. FriReport
  6. SatQuiet
  7. SunQuiet

Messages in the next 7 days

5

in the example week, Wednesday needs action

Quiet days

2

without a message

Confirmed by you

0

changes to the settings

Simple rules recognise fixed phrases in German and English here. No language model is involved, and your input does not leave the browser. The demo makes no claim about recognition quality in our own operations.

How the setup works Reply to the report yourself. The system reads back what it understood, and only after your yes does the coming week change. Switching everything off needs explicit approval.
Scope and limitations

Not part of this demo: This does not make any claim about natural-language recognition quality or about specific recipients and mailboxes.

The setup abstracts a reply channel from our own operations.

Tested in our own operations
We run these three processes ourselves. What is shown are roles, decisions, and reply channels. Recipients, run intervals, operational figures, and content from customer systems are deliberately left out.

Back to the Labor overview →

Which notification really needs to reach someone at your organization?

We work out together which states are relevant, who should act, and what information is actually needed for that.

Schedule a conversation