Optimize AI code assurance portfolio

What it's for

Shows exactly where deeper review, property testing, formal verification or canaries earn their cost—and which AI-code risks remain in the tail.

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.

{
  "assurance_options": [
    {
      "change_stratum_id": "core-service-change",
      "common_defect_detection_probability_scenarios": [
        0.2,
        0.1
      ],
      "defect_detection_sensitivity_scenarios": [
        0.4,
        0.3
      ],
      "delay_hours_per_reviewed_change_scenarios": [
        0.1,
        0.1
      ],
      "dependency_option_ids": [],
      "evidence_verified": true,
      "exclusion_option_ids": [],
      "false_positive_probability_scenarios": [
        0.02,
        0.03
      ],
      "id": "standard",
      "implementation_cost": 0,
      "is_current_state": true,
      "policy_type": "standard_review",
      "remediation_success_probability_scenarios": [
        0.8,
        0.7
      ],
      "resource_demands": {
        "assurance-days": 0
      },
      "review_cost_per_change_scenarios": [
        10,
        10
      ],
      "review_fraction": 0.1,
      "satisfied_control_ids": [
        "review-evidence"
      ],
      "value_retention_fraction": 1
    },

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.

{
  "assumptions": [
    "Defect priors use complete mature aggregate change strata; scenario multipliers preserve common futures and option sensitivity/remediation/false-positive effects come from prospective validation.",
    "Review samples are exchangeable within stratum, common component defects share one event and unique component value is counted once after the union of selected detection controls.",
    "Review, testing, delay, false-positive, remediation, implementation and capacity costs share one finance perimeter and horizon."
  ],
  "assurance_outcomes": {
    "core-service-change": {
      "escape_probability_by_scenario": [
        0.301,
        0.5654
      ],
      "expected_escaped_defects": 0.503
    }
  },
  "decision": "execute_aggregate_ai_code_assurance_portfolio",
  "economics": {
    "expected_escaped_defect_loss": 9592.5,
    "expected_net_value": 9332.55,
    "expected_net_value_gain_vs_current": 27236.4,
    "implementation_cost": 100,
    "loss_and_assurance_cost_conditional_value_at_risk": 110176,
    "loss_and_assurance_cost_value_at_risk": 80540,
    "risk_adjusted_score": -18211.45
  },
  "limitations": [
    "Beta-binomial prevalence and within-stratum exchangeability may miss clustered defects; simulation error and beam search uncertainty are disclosed rather than treated as exactness.",
    "The portfolio allocates aggregate assurance policy, not named reviewers, and cannot infer AI authorship, individual code quality, productivity, misconduct or employment suitability."
  ],
  "method": "bayesian_beta_binomial_ai_code_assurance_cvar_portfolio_v1",
  "pareto_frontier": [
    {
      "expected_net_value": 9332.55,
      "implementation_cost": 100,
      "loss_and_assurance_cost_cvar": 110176,
      "option_ids": [
        "property"
      ]
    },
    {
      "expected_net_value": -17903.85,
      "implementation_cost": 0,
      "loss_and_assurance_cost_cvar": 327632,
      "option_ids": [

How it works

Constrained optimization — Pick the best feasible option under real limits — budget, headcount, dependencies, capacity — rather than ranking a list and hoping it fits.

1. 1 Draw locally calibrated defect prevalence and scenario-conditioned counts per aggregate change stratum, then simulate review, detection, remediation, false positives, delay and escaped defects for every policy.
2. 2 Price direct defect loss and each shared component's unique value once using joint detection survival, preserving common scenarios across the entire portfolio.
3. 3 Reject policy sets that breach escape, evidence, control, dependency, exclusion, budget or resource gates; select expected net value minus loss CVaR by exact enumeration or a disclosed deterministic beam and expose the Pareto frontier.

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": "choose standard expert pair property formal" }
  → finds "optimize_ai_code_assurance_portfolio"

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

gitrevio_capability_run
  { "capability_id": "optimize_ai_code_assurance_portfolio", "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 AI cost, routing & return →