Identify the essential work.
Choose which records need to be available and which actions must continue during an interruption. Some tasks can wait; others may need a separate offline design.
Define what staff should see when the connection drops.
A website that reads well on a phone does not automatically support work without a connection. Plan that behavior around the actual task.
Distinguish local work from a confirmed update.
Staff should understand whether an entry is saved on the device, waiting to send or confirmed in the shared system. Agree on what happens if the device is lost or the page closes.
Plan repeat and conflict handling.
Define how repeated submissions and changes from two people are reviewed. Reconnection should not quietly overwrite a record the team needs to check.
Test the actual conditions.
Use representative devices and fictional records to try interruption, recovery and an unresolved conflict. Confirm what reaches the shared staff record and who handles a failure.