MAM: Completion & Opinion
ISA: ISA 200, 700
Module: operations.audit.reviewai
Mission
This programme is an ABI module with its own agent and Nexus app. Production tooling for each sub-solution is not built yet. The list below is the mission contract from the digested catalog.
Sub-solutions (4)
REVIEWAI.AtlasCheckAtlas File Compliance Check (4 tasks) · Agentic LLM + Rule EngineREVIEWAI.WPCheckWorking Paper Completeness Checker (4 tasks) · Agentic LLM + Rule EngineREVIEWAI.FinalPARFinal Analytical Review Assistant (4 tasks) · Predictive ML + Agentic LLMREVIEWAI.OpinionAssistOpinion Formation Assistant (4 tasks) · Agentic LLM + Benchmark ML
Task-skill pairs (16)
| ID | Sub-solution | Task | Skill |
|---|---|---|---|
| RA-01 | REVIEWAI.AtlasCheck | Verify that all mandatory Atlas fields are populated and signed off per MAM requirements | MandatoryFieldCheckerSkill |
| RA-02 | REVIEWAI.AtlasCheck | Cross-check consistency within Atlas: does the risk assessment align with the response design? Does the response design address all significant risks? | AtlasCoherenceValidatorSkill |
| RA-03 | REVIEWAI.AtlasCheck | Verify that each significant risk has a documented specific response and an interim/final conclusion | SignificantRiskVerifierSkill |
| RA-04 | REVIEWAI.AtlasCheck | Confirm all required approval sign-offs are present (team, manager, partner, EQCR if required) | SignOffCompletenessAuditorSkill |
| RW-01 | REVIEWAI.WPCheck | Scan all WPs for open (uncleared) review notes and report their ageing | ReviewNoteScannerSkill |
| RW-02 | REVIEWAI.WPCheck | Verify that every identified risk has a corresponding completed procedure and documented conclusion in the WPs | RiskToWPCoverageSkill |
| RW-03 | REVIEWAI.WPCheck | Flag WP sections that contain conclusions but lack sufficient documented supporting evidence | MissingEvidenceFlaggerSkill |
| RW-04 | REVIEWAI.WPCheck | Verify file closure requirements (all WPs signed off, review notes cleared, uncorrected misstatement schedule complete) | FileClosureCheckerSkill |
| RF-01 | REVIEWAI.FinalPAR | Build expected final FS values using models from KARMA-PAR updated with any post-fieldwork information | FinalExpectationBuilderSkill |
| RF-02 | REVIEWAI.FinalPAR | Identify and quantify residual unexplained differences between final FS and expectations | DeviationAnalyserSkill |
| RF-03 | REVIEWAI.FinalPAR | Assess whether unexplained differences, individually or in aggregate, require additional audit procedures | ResidualRiskAssessorSkill |
| RF-04 | REVIEWAI.FinalPAR | Generate a draft final analytical review WP memo | FinalPARNarratorSkill |
| RO-01 | REVIEWAI.OpinionAssist | Consolidate all identified and uncorrected misstatements and compare aggregate to materiality thresholds | MisstatementAggregatorSkill |
| RO-02 | REVIEWAI.OpinionAssist | Apply ISA 705/706 decision logic to recommend opinion type (unmodified, qualified, adverse, disclaimer, EOM, OM) | OpinionTypeRecommenderSkill |
| RO-03 | REVIEWAI.OpinionAssist | Propose Key Audit Matters (KAM) candidates for partner consideration | KAMSelectorSkill |
| RO-04 | REVIEWAI.OpinionAssist | Draft the full auditor's report in the applicable format (ISA 700/705/706/701) | ReportDrafterSkill |