Capital Allocation Optimizer GPT — CustomGPT

This CustomGPT brings the capital allocation framework into a conversational interface. You give it initiative data, it applies the scoring rubric, challenges weak assumptions, and outputs a ranked portfolio with IC-ready commentary. It operates as a disciplined investment committee analyst — not a deal promoter.

• Scores initiatives across five weighted dimensions: Value Creation (30%), Strategic Alignment (20%), Balance Sheet Impact (15%), Risk and Execution (20%), and Confidence and Evidence Quality (15%)
• Applies the same scoring rubric every time — decisions are traceable to specific Initiative Card fields, suitable for IC minutes and board packs
• Flags missing or low-confidence inputs before scoring — no fabricated numbers, no false precision
• Drafts board-ready portfolio summaries with ranked initiatives, trade-off commentary, and constraint checks

Best For- CFOs, FP&A directors, and investment committee members who want AI-assisted capital analysis inside ChatGPT

Capital Allocation Data Ingestor — Claude Skill

This Claude Skill reads raw financial documents and maps the data into Initiative Card fields in your Excel workbook. It handles capex forms, business cases, CIMs, and deal models — across PDFs and Excel files. It appends new initiative rows only and generates a change log with source references every time it runs.

• Extracts 40+ financial fields per initiative: NPV, IRR, EBITDA impact, payback, FCF projections, balance sheet effects, and risk flags
• Assigns data confidence scores (1-5) to each extracted field — so you know which numbers need human review before scoring
• Generates a versioned change log after every run, recording what was added, from which source document, and at what confidence level
• Never modifies existing initiative data — strict append-only logic protects decisions already made in prior cycles

Best For- FP&A teams and finance controllers who intake multiple capital requests each planning cycle and need structured, audit-ready data without manual re-entry

Capital Allocation Optimizer Workbook — Excel Model

The Excel workbook is the system's single source of record. It holds every initiative — past and present — and carries them forward across planning cycles. The Claude Skill populates it. The CustomGPT scores from it. The SOP governs how it's used. Nothing gets decided outside it.

• 10 structured tabs: Instructions, Portfolio Constraints, Initiative Cards, Scoring, Pivot by Category, Pivot by Priority, Pivot by Objective, Dashboard, Assumptions Log, Data Lists
• Initiative Cards capture 40+ fields per initiative — identity, economics, scenarios, balance sheet impact, risk, and decision status
• Scoring tab applies weighted criteria consistently across all initiatives in the current cycle
• Assumptions Log preserves all inputs and rationale across cycles — so decisions made 18 months ago are still traceable

Best For- CFOs and FP&A directors running annual or quarterly capital review cycles across multiple business units or deal types

Capital Allocation Knowledge Base — Reference File

This PDF is uploaded directly to the CustomGPT as its persistent knowledge file. It defines the scoring rubric, dimension weightings, guardrails, and interpretation rules the GPT applies to every initiative. Because it's stable, the GPT gives consistent scoring logic across every cycle — regardless of who runs it or what type of deal is being evaluated.

• Defines all five scoring dimensions with full criteria, score anchors (1, 3, 5), and weight percentages
• Establishes guardrails — no initiative proceeds without a downside case, balance sheet impact, and a clear do-nothing consequence
• Scoring tab applies weighted criteria consistently across all initiatives in the current cycle
• Sets priority band thresholds: Proceed, Proceed with Conditions, Defer, and Reject — with decision rules for each

Best For- Finance teams who want the GPT to apply the same methodology across multiple cycles without re-prompting the scoring logic each time

Capital Allocation SOP — Playbook Guide

The SOP defines the six-step decision cycle your team follows every time a capital review runs. It covers what to do at each step, who does it, what inputs are needed, and what the outputs are. It's designed so any finance team member can run the process correctly without needing the person who built the system.

• Six steps: Define Decision Context, Populate Initiative Cards, Auto-Populate via Claude Skill, Score and Rank via CustomGPT, Construct and Stress-Test the Portfolio, Decide and Document
• Defines portfolio constraints to lock before scoring: capital budget, liquidity floor, leverage limits, hurdle rates, and covenant considerations
• Clarifies the separation of responsibilities: Excel as system of record, Claude Skill for ingestion, CustomGPT for decision support, CFO for final authority
• Covers governance: escalation triggers, fast-track conditions, and minimum data quality gates before any initiative enters scoring

Best For- Finance teams deploying the system for the first time, and CFOs who want a documented, auditable process for board or IC purposes

Sample Output — Everest Pipes Worked Example

This is a fully worked example showing the system in action for a fictional manufacturing company, Everest Pipes. The sample Excel workbook shows completed Initiative Cards and scored outputs generated by the CustomGPT. The accompanying PDFs show the raw input documents — a capital investment request and an acquisition info memorandum — that the Claude Skill ingested to produce them.

• Fully populated Excel workbook with real-format initiative data, scoring outputs, and dashboard view — generated by the CustomGPT from Everest Pipes source documents
• Two sample input documents: a capex request (Everest Pipes) and an acquisition info memorandum (SteelTech) — the same document types the Claude Skill is designed to process
• Sample change log showing exactly what the Claude Skill extracted, which fields it mapped, and what confidence scores it assigned after ingesting the Everest Pipes documents

Best For- Anyone evaluating the system before their first live cycle — use this to understand what good output looks like before running your own data through the tools