Optimize AI workflow design portfolio

What it's for

Moves agent optimization from model shopping to whole-workflow design: which graph of models, tools, controls and reviews creates the most value without hiding shared bottlenecks or downside?

What you give it

Every tool is a pure function: it computes from what you pass it and reads nothing on its own. Inputs split into evidence you supply from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.

{
  "design_options": [
    {
      "available_scenario_ids": [
        "base",
        "stress"
      ],
      "cost_per_execution_scenarios": [
        0.1,
        0.1
      ],
      "dependency_option_ids": [],
      "engineering_capacity_units": 0,
      "evidence_verified": true,
      "exclusion_option_ids": [],
      "id": "current",
      "implementation_cost": 0,
      "is_current_state": true,
      "p95_trace_latency_ms_scenarios": [
        150,
        180
      ],
      "resource_demand_scenarios": {
        "model-calls": [
          1000,
          1200
        ],
        "review-hours": [
          0,
          0
        ]
      },
      "satisfied_control_ids": [
        "audit-log"
      ],
      "success_probability_scenarios": [
        0.85,
        0.82
      ],
      "workflow_id": "support"
    },
    {
      "available_scenario_ids": [
        "base",

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": [
    "Every design is prospectively validated against the same workflow outcome, horizon and scenario perimeter; current state is explicit and alternatives include all operating and implementation effects.",
    "Required controls, minimum success, maximum latency and scenario availability are hard local constraints. Shared model, tool, review or other capacities are aggregated bottom-up before selection.",
    "Expected value and regret compare only declared designs; exactness covers the finite option space and beam mode is a disclosed feasible heuristic.",
    "Selection is not deployment, procurement, privacy/security/legal approval, a model/tool SLA, or a judgment about a provider, team or person."
  ],
  "baseline_current_state": {
    "expected_net_value": 534.72,
    "expected_operating_cost": 116,
    "option_ids": [
      "current"
    ]
  },
  "constraints": {
    "engineering_capacity_units": 10,
    "implementation_budget": 100,
    "maximum_cvar_economic_regret": null,
    "maximum_expected_economic_regret": null,
    "minimum_expected_portfolio_net_value": null,
    "tail_probability": 0.1
  },
  "decision": "ai_workflow_design_portfolio_supported",
  "failed_gates": [],
  "method": "governed_ai_workflow_design_multiple_choice_portfolio_v1",
  "option_diagnostics": [
    {
      "eligible": true,
      "failed_gates": [],
      "option_id": "current",
      "workflow_id": "support"
    },
    {
      "eligible": true,
      "failed_gates": [],
      "option_id": "redesigned",
      "workflow_id": "support"
    }
  ],
  "resource_diagnostics": [
    {
      "capacity_by_scenario": [
        2000,
        2000

How it works

Network & dependency analysis — Trace how load, failure and knowledge propagate through a graph of teams, services or components.

1. 1 Compute scenario operating cost, success-adjusted value and declared model/tool/review resource loads for every current and alternative workflow graph.
2. 2 Reject options missing controls, minimum success, latency or availability before economics; then enforce cross-workflow dependencies/exclusions, shared capacities, implementation cash and engineering capacity.
3. 3 Enumerate the finite multiple-choice portfolio exactly when tractable or use a disclosed capacity-pruned beam, optimizing expected value minus CVaR regret and exposing the Pareto frontier plus scenario/resource diagnostics.

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 20 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 ai workflow graph" }
  → finds "optimize_ai_workflow_design_portfolio"

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

gitrevio_capability_run
  { "capability_id": "optimize_ai_workflow_design_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 →