Optimize AI capability resilience portfolio

What it's for

Turns AI deskilling anxiety into an investable resilience frontier: exactly which drills, rotations, dual-running or redesign actions preserve the most operating value under scarce time and money.

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.

{
  "capability_classes": [
    {
      "ai_provider_group_id": "agent-provider",
      "business_value_scenarios": [
        20000,
        20000
      ],
      "fallback_unit_count": 10,
      "id": "deployment",
      "loss_per_shortfall_unit_scenarios": [
        5000,
        20000
      ],
      "maximum_activation_time_hours": 4,
      "maximum_shortfall_probability": 0.9,
      "required_control_ids": [
        "observed-unaided-drill"
      ],
      "required_fallback_units": 6
    }
  ],
  "implementation_budget": 100,
  "provider_groups": [
    {
      "id": "agent-provider",
      "unique_value_at_risk_scenarios": [
        20000,
        100000
      ]
    }
  ],
  "resilience_options": [
    {
      "activation_time_hours_scenarios": [
        3,
        4
      ],
      "capability_class_id": "deployment",
      "dependency_option_ids": [],
      "evidence_verified": true,
      "exclusion_option_ids": [],
      "id": "monitor",
      "implementation_cost": 0,

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": [
    "Each option's scenario readiness probability is prospectively validated for a fungible aggregate capability unit under a real AI-unavailable exercise.",
    "Conditional unit readiness is binomial within a capability class; common provider outage states and unique provider value are priced once across dependent classes.",
    "Drills, rotations, cross-training, dual running and fallback redesign include complete value retention, operating cost, implementation, control, dependency and capacity effects."
  ],
  "decision": "execute_ai_capability_resilience_portfolio",
  "economics": {
    "expected_loss": 21.8331,
    "expected_net_value": 19178.1669,
    "expected_net_value_gain_vs_current": 8654.5791,
    "implementation_cost": 100,
    "loss_conditional_value_at_risk": 108.8443,
    "loss_value_at_risk": 108.8443,
    "risk_adjusted_score": 19150.9558
  },
  "fallback_readiness": {
    "deployment": {
      "expected_unit_shortfall_by_scenario": [
        0.0001,
        0.0018
      ],
      "shortfall_probability_by_scenario": [
        0.0001,
        0.0016
      ]
    }
  },
  "limitations": [
    "Correlated capability-unit failures beyond the represented provider state can understate shortfall; exact optimization proves only the supplied finite model and beam mode does not prove global optimality.",
    "Selection governs aggregate capability classes and does not authorize named-person monitoring, performance scoring, discipline, security accusations or employment decisions."
  ],
  "method": "multiple_choice_ai_capability_binomial_readiness_cvar_v1",
  "pareto_frontier": [
    {
      "expected_net_value": 19178.1669,
      "implementation_cost": 100,
      "loss_cvar": 108.8443,
      "option_ids": [
        "quarterly-drill"
      ]
    },
    {
      "expected_net_value": 10523.5877,

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 Convert each prospectively validated option's conditional unit-readiness probability into exact binomial shortfall probability and expected missing units for every scenario.
2. 2 Reject options that breach shortfall, activation, control, dependency, exclusion, budget, evidence or resource constraints before economic ranking.
3. 3 Price direct shortfall and common-provider unique value once through joint survival, then maximize expected net value minus loss CVaR with exact enumeration or a disclosed deterministic beam.

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 unaided drills work rotations crosstraining" }
  → finds "optimize_ai_capability_resilience_portfolio"

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

gitrevio_capability_run
  { "capability_id": "optimize_ai_capability_resilience_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 →