Course Corrections

Course Corrections

Course Corrections

Course Corrections

or, Why service models must learn from the interactions they produce

or, Why service models must learn from the interactions they produce

or, Why service models must learn from the interactions they produce

or, Why service models must learn from the interactions they produce

Role

Service Design & UX Lead

Discipline

Service Design · UX Design

Focus

Work allocation for Immigration and Asylum

Approach

Story mapping, task modelling, prototyping and usability testing

I led service design and UX for Work Allocation—helping tribunal judges and caseworkers carry out their responsibilities.

Work Allocation triangle with Access Management at the top, Task management at lower left and Case allocation at lower right

Access Management

I translated blueprints and user stories into testable journeys, scenarios, and interactions, then carried what they revealed back into the service model.

End-to-end Immigration and Asylum appeal journey

Making the Service Real

The high-level service had been mapped and approved. Making it tangible produced a different kind of knowledge.

I translated the service model into detailed journeys, role maps, task scenarios, and GOV.UK-based prototypes.

Workshops established the broad direction. Testing the detail revealed consequences that the earlier artefacts could not.

The learning cycle connects the service model, user journeys, scenario maps, prototypes and feedback

Working artefacts did not simply represent the service. They helped the service understand itself.

Decision Time

Work Allocation sought to move cases through the courts more efficiently by changing how judicial work was distributed.

Judges were spending time on routine “boxwork”. Tribunal caseworkers were given extended powers to find, own, and complete more of it—preserving judicial time for complex decisions only judges could make.

H-04 Venn

Routine case management and distributed tasks moved towards tribunal caseworkers.

H-05 Task Push

Judicial attention was reserved for the points at which judicial authority was required.

Lost in Translation

Testing a routine task-assignment flow exposed a distinction the service model had lost: the role that normally owned a task was not the only role allowed to act.

One agreed story allowed a leadership judge to assign judicial work to an appropriate judge. We translated it into a short task-assignment flow, then tested the interaction with judges.

H-06 User story

The screens did what the story asked. Testing revealed that the story was too narrow.

The user story captured what a leadership judge should normally do: assign judicial work to another judge. It did not capture the wider authority salaried judges retained outside that BAU route.

As judges, we need to ensure that we maintain control of the tasks delegated to tribunal caseworkers on our behalf.

Delegation should not remove judicial authority.

Several applications might occur on the day of a hearing that we need to be able to complete without any support.

Urgent work could not always follow the BAU route.

The service had modelled what judges should usually do as the limit of what they could do.

Feedback Loop

What testing revealed travelled back through the service: roles were remapped, user stories were rewritten, All Work was redesigned, and technical work already considered complete was reopened.

The target service used clearly defined case roles to move work through the court efficiently. These established who should normally receive each task.

In practice, people retained the wider authority of their organisational roles.

H-07 Authority

The service had confused what someone should usually do with everything they should be allowed to do.

Correcting it meant retracing earlier decisions through two taxonomically distinct lenses.

Case role describes the work a judge normally does; organisational role describes the work they are allowed to do

The Impact

All Work, Reworked

Once case role became the default lens rather than the boundary of access, almost everything in All Work had to be reconsidered—especially its task view.

We carried the corrected model back into the service definition. New user stories described how people should work across role categories, rather than treating each category as a separate silo.

Testing changed our understanding. The revised stories changed what the service asked the technology to do.

H-09 Feedback

The new stories sent us back through All Work—remapping how roles shaped it and redesigning work already considered done.

Remapped roles, revised stories and redesigned All Work task views

The service model had learned from the interaction it produced.

The Context

The Wider Work

One snapshot from a much broader body of service design and UX work on the Reform project.

Work Allocation within the wider HMCTS Reform programme