01A service request people can follow
Map the request from submission through scheduling, field updates and closure. Decide which details a customer must supply and which the coordinator can add later. Show the difference between an accepted request, a scheduled visit and completed work. Design a useful view for each participant rather than exposing the entire internal record to everyone. A prototype can test whether users understand the next action before the team builds the portal.
02An application that respects existing records
Identify where customer, asset, work-order and inventory records belong. Define the supported interfaces, field mappings and allowed updates before duplicating data into the new application. Scope systems and API integration around a concrete event, such as an approved service request becoming a work order. Decide how the portal handles an unavailable destination, an unfamiliar asset number or a repeated submission. The receiving system should retain a clear record of what was accepted.
03Field screens with recoverable work
Discuss the devices and connection conditions the field team actually uses. A technician may need to save notes, attach photographs or pause a visit while waiting for a part. Specify whether drafts stay on the device, how uploads resume and what happens when two people update the same job. Use explicit states such as saved locally, waiting to send and confirmed by the server. Test interrupted sessions with representative files and devices.
04Permissions and a release the team can maintain
Separate customer access, dispatch permissions and administrative changes. Apply those rules to the application interfaces as well as the screens. Plan account removal, ownership of attachments and the record of consequential changes. For an existing application, rehearse any data migration on a copy and reconcile important records before release. Agree who owns the code, deployment process and support queue so routine changes do not depend on an undocumented handoff.