Skip to main content

9. User Interaction Model

This section defines how users interact with the DXP to manage, process, and complete work across large sets of records. The interaction model is designed for throughput, clarity, and progressive depth.

9.1 Work-Oriented User Journeys​

User journeys are organised around work, not screens or systems.

A work-oriented journey:

  • Starts from an explicit user goal (e.g. review, approve, resolve)
  • Operates over a collection of items
  • Moves from overview to action to completion

Journeys are:

  • Non-linear and interruptible
  • State-aware but not state-dependent
  • Resilient to partial data or degraded execution

The DXP supports users moving fluidly between overview and detail without losing context.

9.2 Work Aggregation and Queues​

The primary interaction surface is a work aggregation view.

Work aggregation:

  • Collects items from one or more capabilities
  • Presents them as queues or lists
  • Reflects current status, priority, and confidence

Queues:

  • Are defined declaratively
  • Support role- and context-based visibility
  • Do not embed workflow logic

This enables scalable handling of large volumes of work without forcing rigid processes.

9.3 Filtering, Ranking, and Focus​

Users refine work using filtering and ranking mechanisms.

Filtering:

  • Applies deterministic criteria
  • Is fast, predictable, and transparent

Ranking:

  • Orders work items based on declared signals
  • May be deterministic or AI-assisted
  • Always exposes ranking rationale when AI-assisted

Focus mechanisms:

  • Allow users to narrow attention to high-value or high-risk items
  • Preserve user intent across navigation and refresh

This prevents cognitive overload in high-volume scenarios.

9.4 Batch Actions​

The DXP supports batch actions as first-class interactions.

Batch actions:

  • Operate over selected sets of work items
  • Invoke capabilities with aggregated inputs
  • Provide preview and confirmation steps

Batch execution:

  • Is observable and cancellable
  • Supports partial success and rollback where applicable
  • Reports confidence and errors per item

This enables throughput without sacrificing control.

9.5 Deep-Dive Interaction​

From aggregated views, users can drill into individual items.

Deep-dive views:

  • Present complete context for a single item
  • Surface provenance and confidence clearly
  • Support focused, atomic actions

Deep-dive interaction:

  • Does not trap users
  • Preserves return-to-work context
  • Allows escalation or exception handling

This ensures depth without disrupting flow.