Optimize AI knowledge refresh portfolio

What it's for

Finds the economical refresh frontier across an entire AI estate, recognizing that one policy page or runbook may ground many assistants and should be refreshed—and paid for—once.

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.

{
  "ai_applications": [
    {
      "evidence_verified": true,
      "id": "support-copilot",
      "loss_per_stale_answer_scenarios": [
        100,
        200
      ],
      "maximum_expected_stale_loss": 100000,
      "minimum_grounding_probability": 0.5,
      "query_volume_scenarios": [
        1000,
        1200
      ],
      "required_source_ids": [
        "payments-runbook"
      ],
      "value_per_grounded_answer_scenarios": [
        5,
        5
      ]
    }
  ],
  "implementation_budget": 10000,
  "knowledge_sources": [
    {
      "daily_change_hazard_scenarios": [
        0.01,
        0.02
      ],
      "evidence_verified": true,
      "id": "payments-runbook",
      "required_control_ids": [
        "access"
      ]
    }
  ],
  "refresh_options": [
    {
      "available_scenario_ids": [
        "base",
        "stress"
      ],

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.

{
  "application_diagnostics": [
    {
      "application_id": "support-copilot",
      "expected_grounding_probability": 0.9762,
      "expected_stale_loss": 4047.6584,
      "minimum_grounding_probability": 0.5
    }
  ],
  "assumptions": [
    "Each immutable knowledge source is refreshed once for all dependent AI applications; shared-source benefits, costs, controls and capacity are never counted once per consuming application.",
    "Periodic-refresh freshness follows a renewal model with Poisson content change and prospectively tested refresh success; application grounding requires every declared source to be fresh.",
    "Expected value and CVaR regret use one coherent scenario order. Required controls, grounding thresholds, application loss limits, relations, budget and shared resources are hard gates.",
    "Optimization is not deployment, content-access authorization, factual certification, legal/compliance advice, or a judgment about an author, employee, vendor, team, or country."
  ],
  "baseline_current_state": {
    "expected_net_value": -33192.2349,
    "expected_stale_loss": 36920.3177,
    "option_ids": [
      "monthly"
    ]
  },
  "constraints": {
    "implementation_budget": 10000,
    "maximum_cvar_economic_regret": null,
    "maximum_expected_economic_regret": null,
    "risk_aversion": 0.25,
    "tail_probability": 0.05
  },
  "decision": "ai_knowledge_refresh_portfolio_supported",
  "failed_gates": [],
  "method": "shared_source_renewal_refresh_portfolio_cvar_regret_v1",
  "portfolio_pareto_frontier": [
    {
      "economic_regret_cvar": 82908.719,
      "expected_net_value": -33192.2349,
      "implementation_cost": 0
    },
    {
      "economic_regret_cvar": 0,
      "expected_net_value": 501.283,
      "implementation_cost": 5000
    }
  ],

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 source change hazards, refresh intervals and tested refresh success into scenario freshness with a periodic-renewal model.
2. 2 Propagate each source choice once through all dependent applications, where grounding requires every declared source; subtract stale loss and shared operating cost without double counting.
3. 3 Enforce controls, scenario availability, dependencies, exclusions, application grounding/loss limits, implementation budget and shared capacity; solve exactly or by disclosed beam and rank expected value minus CVaR regret.

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": "select one governed periodic refresh policy" }
  → finds "optimize_ai_knowledge_refresh_portfolio"

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

gitrevio_capability_run
  { "capability_id": "optimize_ai_knowledge_refresh_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 →