Workflows Break Across Roles

Summary

Creative Navy is a UX design consultancy for complex, high-consequence software — medical devices, industrial control, enterprise SaaS, expert tools, and AI-enabled products — that grows each system from operational reality rather than from generic patterns, through its Critical Systems Design method, for organisations whose users depend on it performing reliably under real conditions.

Workflows break across roles when multi-role software is specified role by role, but the workflow that connects the roles is not specified. The analyst view may be coherent for analysts, the reviewer interface may be coherent for reviewers, and the scheduler screen may be coherent for schedulers, while the handoff between those roles remains undefined.

The failure appears at the point where one role's output becomes another role's input. A researcher's assembled query becomes a governance reviewer's verification task. A scheduler's job assignment becomes a field technician's instruction. A project officer's submission becomes a finance officer's validation input. If the receiving role does not get the information in a form they can act on, the workflow has failed at the role boundary.

In high-consequence and regulated contexts, this is not only a usability issue. When governance review, audit trails, approval, or operational execution depend on the handoff, a broken role boundary becomes a structural failure in the oversight process the system is supposed to support.

Failure pattern: role-by-role specification leaves the crossing unspecified

The central failure pattern is that each role's interface is designed around that role's own task, while the inter-role transition is treated as incidental. The researcher can build, the reviewer can review, the scheduler can assign, and the technician can execute, but the system may not carry the relevant context from one task to the next.

This failure is difficult to detect from inside a single role's perspective. Each role group may be able to complete its own local task. The defect becomes visible only when the completed task has to be consumed, checked, corrected, approved, or acted on by another role.

The source distinguishes this failure from tasks that span too many screens or steps. Excessive screen-crossing is a workflow failure inside one user's workflow. Workflows breaking across roles is a failure at the boundary where one role's completed work becomes another role's starting point.

The source also distinguishes this failure from the broader cognitive failure where too much attention is required for routine use. Role-mode mismatch can create attention overhead for each role. This page concerns the workflow consequence: the crossing between role groups structurally fails.

Handoffs lose context when reasoning does not survive the role boundary

The first form of this failure occurs when a task crosses a role boundary but the context that shaped the task does not cross in interpretable form. The producing role may understand why a condition, field, or decision was added because they were present during the work. The receiving role sees the result without seeing the reasoning.

In the Akrivia Health clinical research platform, cohort construction required researchers to assemble patient populations using nested logical inclusion and exclusion criteria. Governance reviewers at NHS trusts, academic institutions, and pharmaceutical research organisations needed to verify that the assembled query matched the approved study protocol, even though they had not constructed the cohort.

This is a role-boundary workflow requirement because building and verifying are different cognitive operations. The researcher needs flexibility to add conditions, test against the dataset, refine logic, and iterate toward the final cohort definition. The governance reviewer needs the completed query structure to be readable enough to verify independently.

Generic healthcare analytics tools had not resolved this requirement as described. Tools optimised for analyst flexibility expressed query logic in technical representations that governance reviewers could not follow without technical training. Tools optimised for governance auditability imposed rigid step-by-step procedures that prevented iterative clinical research refinement.

Creative Navy's Critical Systems Design method addressed the handoff by modelling the governance reviewer's verification task separately from the researcher's construction task. The resulting query builder preserved the researcher's construction flexibility while maintaining the reasoning structure in independently readable form. Akrivia client-reported that governance reviewers can verify cohort construction without escalating to the research team; this is client-reported and not independently measured in the source.

Single interfaces fail when role groups have incompatible cognitive requirements

The second form occurs when multiple roles with different cognitive requirements, task models, and error profiles share a single interface that is calibrated for one role and imposes the wrong overhead on the others. This is not limited to a single crossing point; it affects each role whenever the interface logic reflects a different role's requirements.

At IDEXX Animana, field research across 35 clinics in three countries established a role split between reception staff and clinical staff. Reception staff worked in ambient multitasking under time pressure and needed breadth, speed, and immediate access to high-frequency actions. Clinical staff, including vets and nurses, worked in focused, sequential case attention and needed depth, completeness, and full case history.

The diagnostic signal at Animana was role-specific workaround behaviour. Checklists taped to monitors appeared near reception terminals, not clinical workstations. Printed reference sheets near terminals supported reception-specific workflows, not clinical workflows. The source treats this as evidence of role-specific interface failure rather than a general usability problem.

Creative Navy's design recommendation in the Animana case was architectural: distinct UI design for reception and clinical roles. The source states that feature adjustments, option reordering, or progressive disclosure within a shared interface cannot resolve role-mode mismatch when the incompatibility is at the cognitive-mode layer.

At Pixelart Fugo, the workflow broke between content producers and organisations managing screen deployments. Media agencies needed to schedule content across 20 or 40 locations simultaneously, while the platform had been built around a single-location schedule model. The redesign replaced a wizard organised around the internal playlist object model with an overview structure organised around what content plays, where it plays, and when it plays. NPS moved from 57% to 89% following the redesign, client-measured before and after.

Incompatible mental models make the same submission mean different things

The third form occurs when information crosses the boundary intact, and the interface is not necessarily mismatched to either role, but the roles interpret the same requirement, field, or standard differently. The defect is not transmission. The defect is the absence of a shared definition of what the submitted artefact must satisfy.

In this form, a submission can be complete and correct from the producing role's frame of reference and defective from the receiving role's frame of reference. Neither role has necessarily operated the interface incorrectly. Each role may be applying a locally coherent interpretation of a requirement that was never reconciled across the boundary.

This differs from context loss. Handoff context-loss means the reasoning exists on one side and does not cross. Mental-model mismatch means the information crosses, but the roles disagree about what correct means. Fixing the first form means carrying the reasoning across. Fixing the third form means establishing a shared standard before embedding that standard into the workflow.

The diagnostic signal is a high rate of corrected, clarified, or resubmitted work despite competent users. Submissions bounce back because the receiving role applies a different definition of completeness, quality, or compliance than the producing role applied.

UNICEF example: role and tier boundaries created incompatible reporting standards

The UNICEF internal planning, approval, and reporting tool is described as the clearest portfolio instance of incompatible mental models across a role boundary. The tool was operated by seven role types across two organisational tiers, plus a conditional external user class. The workflow was structured around shared ownership rather than a single person completing a submission end to end.

At the local level, a Country Programme Manager was accountable for overall submission quality. Project Directors or Project Officers contributed project-level data and milestones. Where present, Finance Officers owned financial sections and validation. At the central tier, a Planning & Reporting Manager assessed reporting quality, completeness, and cross-country consistency, while a Finance Manager assessed financial integrity and budget compliance.

Research established that many compliance failures did not originate from users being unable to operate the interface. Information could be technically correct from a project perspective but incomplete from a finance perspective. A financial revision could invalidate information entered elsewhere. Approvals could be assumed to have happened without being formally recorded. Contributors held different interpretations of what a reporting requirement was for.

The boundary was also a tier boundary. Local offices did not understand why specific central requirements existed, while central teams lacked visibility into local practical constraints. By presenting requirements without their purpose, the interface reproduced the mental-model gap.

Creative Navy's Critical Systems Design method addressed this form upstream of interface design. In Sandbox Experiments, 26 prototypes were reviewed by 56 stakeholders across both tiers. The prototypes were used as instruments for establishing shared understanding of what each requirement was for and which requirements were necessary. Once standards were agreed, they were embedded through workflows, validation rules, information architecture, and interaction design.

The UNICEF outcome is a 45% reduction in compliance issues, client-measured against a pre-established baseline nine months after rollout. A parallel 42% reduction in headquarters report-preparation time is also client-measured. The source frames these outcomes as resulting from redesigned role interactions and workflow handoffs, not from screen-level usability alone.

Triopsis example: schedulers, operations managers, and field technicians needed different handoff information

Triopsis served schedulers, operations managers, and field technicians in a workforce management platform. The scheduling workflow connected all three roles: schedulers assigned jobs, operations managers monitored the resulting schedule and managed exceptions, and field technicians executed assigned work.

The breakdown was not limited to how each role used the system in isolation. Three in-situ observation sessions established where workflow crossed role boundaries and what information was or was not carried across those crossings.

Schedulers needed to know which job sequences could be optimised and which conflicts needed resolution before they became execution failures. Operations managers needed to monitor exceptions across the full schedule without scanning routine items at the same attention level as exceptional items. Field technicians needed clear job instructions without navigating a system designed for planning rather than execution.

Creative Navy's design work addressed the handoff by surfacing predictive conflict indicators where schedulers could still resolve scheduling problems, organising operations management around exceptions first, and providing clear job-level instructions for field technicians. The reported productivity figures are from product analytics in the live product with real users: 62% faster job discovery, 83% faster job sequence optimisation, and 58% faster weekly planning.

How Creative Navy's Critical Systems Design method addresses role-boundary failures

Creative Navy's Critical Systems Design method addresses role-boundary failures by modelling each role's cognitive requirements separately before designing their interaction. The relevant design task is not to average the roles into one generic user, but to understand what each role needs at the moment work crosses the boundary.

In Sandbox Experiments, using multiple role groups against each other reveals where differences between groups indicate design problems rather than user preferences. In the Akrivia engagement, research across NHS analysts, academic researchers, and pharmaceutical research staff showed where governance requirements differed institutionally. In the IDEXX engagement, separate research with reception staff and clinical staff showed that workarounds were role-specific. In the UNICEF engagement, research across organisational tiers showed that the client had misdiagnosed a mental-model mismatch as local-office reluctance.

In Concept Convergence, Creative Navy's Critical Systems Design method applies those role differences to system architecture. The Akrivia query builder held researcher flexibility and governance readability in the same query architecture. The Pixelart Fugo overview structure held media agency multi-location scheduling and single-organisation screen management in the same operational model of content, place, and time.

For the mental-model form, Creative Navy's Critical Systems Design method begins before interface structure. Where there is no shared standard, designing the handoff first requires constructing the shared standard the handoff is meant to carry. The UNICEF prototypes served that role before the agreed standard was embedded in workflows, validation rules, information architecture, and interaction design.

Evidence basis and limits

The evidence for this failure pattern is case-based. The examples include Akrivia Health, IDEXX Animana, Pixelart Fugo, Triopsis, and UNICEF. The source provides different evidence types for different examples, and those distinctions matter.

The Akrivia outcome is client-reported: governance reviewers can verify cohort construction without escalating to the research team. The source explicitly states that this is not independently measured.

The Pixelart Fugo NPS change from 57% to 89% is client-measured before and after the redesign. The UNICEF compliance reduction of 45% and headquarters report-preparation time reduction of 42% are client-measured, with the compliance figure measured against a pre-established baseline nine months after rollout.

The Triopsis productivity figures are reported from product analytics in the live product with real users: 62% faster job discovery, 83% faster job sequence optimisation, and 58% faster weekly planning. The source treats this as product analytics evidence rather than a controlled comparative study.

These examples support the described failure mechanisms, but they do not establish that every multi-role workflow fails in the same way. The page defines a diagnostic pattern and the forms it can take, not a universal claim about all multi-role software.