Audit recommendation coherence

What it's for

Gives Gitrevio an internal truth layer: agents can see when several analytical answers genuinely agree, reuse the same evidence, are stale, or answer subtly different questions before presenting one executive recommendation.

What you give it

Inputs split into evidence read from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.

{
  "current_period": 10,
  "recommendations": [
    {
      "action_id": "fund",
      "decision_contract_id": "roadmap-v1",
      "decision_id": "roadmap-funding",
      "evidence_source_ids": [
        "resolved-projects"
      ],
      "function_id": "delivery-forecast",
      "id": "delivery-model",
      "issued_period": 8,
      "scenario_version": "finance-v1",
      "supported": true,
      "valid_until_period": 12,
      "weight": 1
    },
    {
      "action_id": "fund",
      "decision_contract_id": "roadmap-v1",
      "decision_id": "roadmap-funding",
      "evidence_source_ids": [
        "customer-experiment"
      ],
      "function_id": "value-model",
      "id": "value-model",
      "issued_period": 8,
      "scenario_version": "finance-v1",
      "supported": true,
      "valid_until_period": 12,
      "weight": 1
    },
    {
      "action_id": "fund",
      "decision_contract_id": "roadmap-v1",
      "decision_id": "roadmap-funding",
      "evidence_source_ids": [
        "resolved-projects"
      ],
      "function_id": "risk-model",
      "id": "risk-model",
      "issued_period": 8,
      "scenario_version": "finance-v1",

What you get back

This is the actual output of running the example above — computed by the same function the platform calls, not an illustration.

{
  "configuration": {
    "current_period": 10,
    "maximum_incomparable_weight_fraction": 0.1,
    "maximum_lineage_adjusted_conflict_fraction": 0.2,
    "maximum_stale_weight_fraction": 0.1,
    "maximum_unsupported_weight_fraction": 0.1,
    "minimum_aligned_recommendations": 2,
    "minimum_independent_evidence_sources": 2
  },
  "conflicting_recommendation_pairs": [],
  "decision": "represented_recommendations_are_coherent",
  "decision_diagnostics": [
    {
      "aligned_current_supported_count": 3,
      "aligned_decision_contract_id": "roadmap-v1",
      "aligned_scenario_version": "finance-v1",
      "consensus_action_id": "fund",
      "decision": "coherent_recommendations_supported",
      "decision_id": "roadmap-funding",
      "failed_gates": [],
      "incomparable_weight_fraction": 0,
      "independent_evidence_source_count": 2,
      "lineage_adjusted_action_support": {
        "fund": 1
      },
      "lineage_adjusted_conflict_fraction": 0,
      "maximum_pairwise_evidence_jaccard_overlap": 1,
      "raw_action_support": {
        "fund": 1
      },
      "recommendation_count": 3,
      "stale_weight_fraction": 0,
      "unsupported_weight_fraction": 0
    }
  ],
  "guardrails": [
    "Recommendations are comparable only inside the same governed decision contract, scenario version and validity window; mismatched scopes are not votes against one another.",
    "Every unique evidence source receives one provenance vote split across its claimants, so several tools reusing the same lineage cannot manufacture independent consensus.",
    "Coherence means represented analytical advice agrees after declared provenance adjustment. It does not prove the consensus action is causally valid, valuable, safe or authorized."
  ],
  "method": "provenance_adjusted_recommendation_coherence_audit_v1",
  "summary": {
    "action_required_decision_count": 0,

How it works

Statistical audit & measurement — Check whether a number is fit to decide on: coverage, timing, reconciliation, and the gaps a dashboard hides.

1. 1 Group immutable recommendations by governed decision, select the weight-dominant decision-contract and scenario version, and retain only supported recommendations inside their declared validity window.
2. 2 Calculate raw action support, then give every unique evidence source one vote split across recommendations citing it; diagnose lineage-adjusted conflict, stale, unsupported and incomparable weight plus pairwise evidence overlap.
3. 3 Expose a consensus action only when aligned support, independent lineage, conflict, scope, freshness, support and unique-consensus gates all pass.

Before you trust it

Every tool in the catalog ships with the conditions under which its answer is meaningful — and the conditions under which it should abstain instead of guessing.

Calibrating it to your org

Calibration workflow

1. 1 Define the management decision, target outcome, aggregate unit, privacy boundary, cadence, and prediction/intervention horizon for this organization.
2. 2 Build a tenant-scoped historical cohort using only information available before each prediction or decision; preserve zero periods, censoring, assignment probabilities, and unresolved outcomes when the method requires them.
3. 3 Estimate statistical parameters on training history, but obtain costs, utilities, risk tolerance, practical-effect thresholds, capacity, and policy constraints from accountable decision owners.
4. 4 Validate on later time windows or held-out aggregate units at the deployment grain, against a simple baseline and the function-specific validation strategy.
5. 5 Deploy only if the returned decision clears evidence, overlap, calibration, robustness, and guardrail checks; warning, unsupported, schema-gap, and fallback decisions are abstentions.
6. 6 Monitor realized outcomes, data drift, coverage, and decision regret; recalibrate at a governed cadence or after a detected regime/definition change, never merely because a stakeholder dislikes the result.

Call it from your AI

You don't wire up 388 tools in your MCP client. The GitRevio MCP server exposes 18 tools, three of which let an agent search the catalog, read a tool's schema, and run it — so the assistant finds this one on its own.

gitrevio_capabilities_search
  { "q": "audit whether analytical recommendations for the" }
  → finds "audit_recommendation_coherence"

gitrevio_capability_describe
  { "capability_id": "audit_recommendation_coherence" }
  → returns the input schema and agent guidance shown on this page

gitrevio_capability_run
  { "capability_id": "audit_recommendation_coherence", "arguments": { ... } }
  → returns the result shown above

Works in Claude Desktop, Claude Code, Cursor, Cline, Continue.dev, Goose and Aider. See the MCP server.

Related tools

See every tool in Analytics assurance & orchestration →