Context
MIDAS currently captures individual evaluations, envelopes, and audit events, but it does not yet provide a first-class model for tracing escalation paths and end-to-end decision journeys across multiple related evaluations and review steps.
In real-world scenarios, a governed decision may begin with automated evaluation, move into escalation or review, trigger follow-up decisions, and continue across a broader execution path.
Problem
There is currently no first-class way to model or trace an end-to-end decision journey across related evaluations, escalations, and human interventions.
This makes it harder to:
- understand how a decision progressed over time
- trace escalation paths across multiple steps
- connect human review and follow-up decisions into the same journey
- support future journey-level analytics and operational tooling
Proposed solution
Introduce a first-class journey model for MIDAS that can represent escalation paths and end-to-end execution flow across related evaluations.
A journey should allow MIDAS to connect:
- initial evaluation
- escalation events
- review activity
- follow-up decision evaluations
- final outcome path
This should build on run-level linkage rather than treating each evaluation as an isolated event.
Scope
- Define how end-to-end decision journeys are represented
- Track escalation paths across related evaluations
- Link review and follow-up activity into the same journey context
- Expose journey linkage through runtime artefacts and/or APIs
- Preserve backward compatibility for evaluations that do not participate in a broader journey
Acceptance criteria
- A first-class journey or equivalent linkage model exists
- Escalation paths can be traced across related evaluations
- Review and follow-up activity can be linked into the same journey context
- Journey structure is deterministic and serialisable
- Existing single-evaluation flows continue to function without requiring journey context
- Feature is documented
Context
MIDAS currently captures individual evaluations, envelopes, and audit events, but it does not yet provide a first-class model for tracing escalation paths and end-to-end decision journeys across multiple related evaluations and review steps.
In real-world scenarios, a governed decision may begin with automated evaluation, move into escalation or review, trigger follow-up decisions, and continue across a broader execution path.
Problem
There is currently no first-class way to model or trace an end-to-end decision journey across related evaluations, escalations, and human interventions.
This makes it harder to:
Proposed solution
Introduce a first-class journey model for MIDAS that can represent escalation paths and end-to-end execution flow across related evaluations.
A journey should allow MIDAS to connect:
This should build on run-level linkage rather than treating each evaluation as an isolated event.
Scope
Acceptance criteria