aibvf-mcp

The scoring tool your Claude agent calls before it recommends an AI deployment. Four pillars, published benchmarks, deterministic classification of Accelerate, Fix, or Stop, with modelled EUR value, decision confidence, and a specific list of what to do next.

What It Does

Six tools on stdio, each callable from any MCP-compatible agent.

30-Second Install

npm install -g aibvf-mcp

Register with Claude Desktop, Claude Code, or any MCP client:

{
  "mcpServers": {
    "aibvf": { "command": "aibvf-mcp" }
  }
}

Ask your agent: "score a gen2 CX AI initiative for a 400M EUR retailer, traditional readiness, SA 70, FR 50, CE 55, GR 45," and the agent will call score_initiative, return a Fix classification with a concrete gap list, and offer to call recommend_improvements next.

Why This Exists

Agents confidently recommend AI projects with no reference to the business case, no reference to operating-model readiness, and no reference to governance exposure. The scoring belongs in the agent's pre-flight check, not in a slide deck written after the decision.

The protocol is open, the benchmarks cite McKinsey, Gartner, BCG, Deloitte, Forrester, Accenture, ServiceNow, and readiness capture rates come from EY/Oxford and Prosci change-success research.

The Four Pillars

Every initiative is scored on four pillars, 0 to 100, honest self-assessment.

1. Strategic Alignment — how clearly this moves a board-level KPI.
2. Financial Return — strength of the modelled return.
3. Change Enablement — sponsor in place, owner named, change budget funded.
4. Governance Risk — regulatory and reputational exposure. Higher value means more risk.

Rules are deterministic, no network, no dependencies. GR >= 70 or FR <= 20 returns Stop, all four pillars at or above 60 with GR <= 40 returns Accelerate, anything else returns Fix with a specific gap list.

See docs/scoring-formulas.md for every formula and docs/worked-example.md for a full run on a healthcare portfolio.

Example: Scoring an Agentic Healthcare Initiative

import { score, recommendImprovements, calculatePaceLayerDrag } from '@aibvf/core';

const r = score({
  industry: 'healthcare',
  revenue_eur: 800_000_000,
  function: 'cx',
  ai_tier: 'gen3',
  readiness: 'traditional',
  scores: {
    strategic_alignment: 75,
    financial_return:    55,
    change_enablement:   40,
    governance_risk:     55,
  },
});
// { classification: 'Fix', net_low_eur: 23_760_000, net_high_eur: 83_160_000,
//   confidence: 54, applied_modules: ['four_pillar_base',
//   'readiness_capture_traditional', 'healthcare_clinical_validation',
//   'healthcare_regulatory_overhead'], ... }

Same inputs through recommendImprovements return three pillar raises, each with a named action, and project a new decision confidence of 68 with target classification Accelerate. `calculatePaceLayerDrag({ revenue_eur: 800_000_000, ai_tier: 'ge

... View full README on GitHub