Forecast engineering investment benefit realization

What it's for

Gives CTOs, CEOs, and investors a disciplined answer to whether the engineering investment portfolio is converting promised value into realized economics—using the company's own resolved wins and failures instead of activity-based ROI theater.

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_investments": [
    {
      "category": "platform",
      "committed_cost": 300,
      "elapsed_days": 20,
      "id": "current-platform-investment",
      "planned_benefit": 1000,
      "realized_benefit_to_date": 50
    }
  ],
  "historical_investments": [
    {
      "category": "product",
      "id": "resolved-investment-0",
      "planned_benefit": 100,
      "realization_days": 60,
      "realized_benefit": 0
    },
    {
      "category": "platform",
      "id": "resolved-investment-1",
      "planned_benefit": 100,
      "realization_days": 61,
      "realized_benefit": 140
    },
    {
      "category": "product",
      "id": "resolved-investment-2",
      "planned_benefit": 100,
      "realization_days": 62,
      "realized_benefit": 140
    },
    {
      "category": "platform",
      "id": "resolved-investment-3",
      "planned_benefit": 100,
      "realization_days": 63,
      "realized_benefit": 140
    },
    {
      "category": "product",
      "id": "resolved-investment-4",
      "planned_benefit": 100,

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": [
    "Historical planned and realized benefits use a stable finance-owned definition, resolved initiatives are not selectively omitted, and categories are comparable enough for partial pooling.",
    "The hurdle separates zero-benefit investments from positive benefit magnitude; positive benefit multiples and realization lags are approximately lognormal after category pooling.",
    "Organization shock correlation represents shared execution or market conditions across current investments; committed cost is the relevant decision denominator and benefits arriving after the horizon do not support the horizon decision."
  ],
  "configuration": {
    "annual_discount_rate": 0.1,
    "category_prior_strength": 10,
    "horizon_days": 365,
    "maximum_benefit_multiple": 10,
    "minimum_median_roi": 0,
    "minimum_positive_npv_probability": 0.8,
    "organization_shock_correlation": 0.25,
    "simulations": 500
  },
  "decision": "benefit_realization_economically_supported",
  "detail_counts": {
    "current_investments": 1
  },
  "executive_summary": {
    "committed_cost": 300,
    "expected_discounted_eventual_benefit": 1353.6824,
    "expected_discounted_horizon_benefit": 1353.6824,
    "horizon_npv_confidence_interval": [
      970.6625,
      1191.5516
    ],
    "horizon_npv_p10": 993.63,
    "horizon_npv_p50": 1077.7689,
    "horizon_npv_p90": 1171.3997,
    "median_horizon_roi": 3.5926,
    "planned_benefit": 1000,
    "probability_positive_horizon_npv": 0.976
  },
  "interpretation": "This forecasts portfolio benefit realization from locally resolved investment outcomes. It does not turn engineering activity into revenue, prove that an investment caused observed value, or replace finance-owned benefit attribution.",
  "investment_risk": [
    {
      "category": "platform",
      "committed_cost": 300,
      "historical_category_rows": 25,
      "horizon_npv_p10": 993.63,
      "horizon_npv_p50": 1077.7689,
      "horizon_npv_p90": 1171.3997,

How it works

Sequential Bayesian & bandits — Learn while deciding — update beliefs as evidence arrives and choose where the next unit of effort is worth spending.

1. 1 Normalize resolved historical initiatives to final realized-benefit/planned-benefit multiples and realization lags, keeping zero-benefit outcomes in a separate hurdle instead of deleting failed investments.
2. 2 Estimate organization-wide hurdle, log-benefit-multiple, and log-lag distributions, then partially pool each investment category toward that organization evidence using the governed prior strength; unseen categories use a disclosed global prior only.
3. 3 Draw each category's hurdle probability and lognormal locations once per posterior simulation so epistemic uncertainty is shared across its current investments, then simulate capped benefit multiples and remaining lags with organization-level shocks rather than assuming convenient independence.
4. 4 Credit benefit already realized, discount remaining benefit only when it arrives, exclude benefit beyond the governed horizon from horizon NPV, and gate the portfolio on probability of positive NPV plus median ROI while retaining eventual-value context.

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": "forecast whether an engineeringinvestment portfolio will" }
  → finds "forecast_engineering_investment_benefit_realization"

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

gitrevio_capability_run
  { "capability_id": "forecast_engineering_investment_benefit_realization", "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 Measurement integrity →