Skip to main content

Defects, Anti-Patterns & Standards

Logic Insights checks each process for problems at three levels: structural defects that are objectively wrong, anti-patterns that are recurring bad shapes, and standards gaps measured against the frameworks your organization answers to. The first two are found deterministically from the graph. The third scores the process against tunable numeric criteria.

Design defects

Design defects are structural problems that are wrong by construction. They are detected in code, with no interpretation involved, so they are never false alarms of the "the model thinks maybe" kind.

DefectWhat it means and why it matters
Impossible SLAThe stage durations cannot fit inside the declared SLA. The process is set up to miss its target on every instance.
Orphan StageA stage that cannot be reached from the start. Dead configuration that will never execute.
Dead StageA non-terminal stage with no way out. Work that enters it gets stuck.
Missing Start StageNo entry point can be identified. The process has no defined beginning.
Multiple Unconnected Start StagesMore than one disconnected entry point, usually a modeling mistake.
Cycle Without ExitA loop with no path to a terminal stage. Instances can circle forever.
Duplicate TransitionRepeated transitions from the same source, which create ambiguous routing.
Terminal Stage with Outgoing TransitionA stage marked as terminal that still routes onward, a contradiction in the design.
Segregation of Duties ViolationOne actor holds responsibilities that should be separated, a governance and audit risk.

Anti-patterns

Anti-patterns are higher-level shapes that are not strictly invalid but reliably cause pain. Each is detected using a configurable threshold, so you can tune sensitivity to your own tolerance.

Anti-patternWhat triggers itWhy it hurts
God StageOne stage owns a disproportionate share of the total SLAThe whole process rises and falls with a single stage; it is your most likely bottleneck
Spaghetti FlowCyclomatic complexity above a healthy thresholdHard to understand, hard to change safely, and defect-prone
Bottleneck ActorOne actor owns too large a share of the manual stagesKey-person risk; the process stalls when that person is unavailable
SLA BlindspotA large portion of the process has no SLA coverageYou cannot manage what you do not measure; time leaks in the dark
Manual Bottleneck IslandA run of consecutive manual stages with no automationA slow, error-prone stretch that is a prime automation target

The thresholds for each anti-pattern are per-project settings. See Configuration & API.

Standards audit

Logic Insights scores each process Pass, Partial, or Fail against five widely used frameworks. Each framework is evaluated against numeric criteria that you can tune, so the audit reflects your organization's bar rather than a fixed opinion.

FrameworkWhat it checks
ISO 9001Responsible-actor coverage, entry and completion criteria, exception handling, and escalation paths
COBITApproval-stage ratio and control gates, applied to larger processes
LeanAutomation rate, rework index, and measurable completion criteria
Six SigmaSLA-gap ceiling, defect-prevention criteria, and estimated sigma level
ITILDocumentation coverage and problem-management stages

Each control reports its own Pass, Partial, or Fail with the reason, and the framework rolls up to an overall status using configurable Pass and Partial thresholds. This gives compliance and audit functions a concrete, per-control answer to "does this process satisfy the standard, and where does it fall short."

Regulatory mapping and benchmarking

Beyond the five operational standards, the AI layer maps each process to the regulatory frameworks that matter to you (GDPR, SOX, ISO 9001, HIPAA, PCI-DSS, ISO 27001 by default, all configurable) and benchmarks it against industry dimensions. Both are covered in Portfolio & AI Redesign.