https://github.com/sandeepkunkunuru/idea-evaluator
CLI + interactive xlsx for evaluating business ideas: portfolio screening (RICE/WSJF/CD3/EMV), weighted rubric with AI-resilience scoring, AI-native org design, 7 valuation methods, and Slicing-Pie founder equity splits.
https://github.com/sandeepkunkunuru/idea-evaluator
ai-native business-evaluation cap-table claude-code dcf entrepreneurship founder-equity openpyxl org-design python-cli rice-scoring slicing-pie startup-tools valuation wsjf
Last synced: 22 days ago
JSON representation
CLI + interactive xlsx for evaluating business ideas: portfolio screening (RICE/WSJF/CD3/EMV), weighted rubric with AI-resilience scoring, AI-native org design, 7 valuation methods, and Slicing-Pie founder equity splits.
- Host: GitHub
- URL: https://github.com/sandeepkunkunuru/idea-evaluator
- Owner: sandeepkunkunuru
- License: mit
- Created: 2026-05-17T13:01:02.000Z (about 1 month ago)
- Default Branch: main
- Last Pushed: 2026-05-17T14:00:04.000Z (about 1 month ago)
- Last Synced: 2026-05-17T16:28:48.617Z (about 1 month ago)
- Topics: ai-native, business-evaluation, cap-table, claude-code, dcf, entrepreneurship, founder-equity, openpyxl, org-design, python-cli, rice-scoring, slicing-pie, startup-tools, valuation, wsjf
- Language: Python
- Size: 180 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# idea-evaluator
CLI + interactive xlsx for evaluating business ideas end-to-end:
- **Screen** a portfolio of ideas via Modified RICE, WSJF, CD3, EMV
- **Rubric-score** a single idea on 10 weighted dimensions (incl. AI-resilience)
- **Design an AI-native org structure** with role-by-role automation and cost projections
- **Value** the idea seven ways: DCF, Comparables, Berkus, Scorecard (Payne), VC method, First Chicago, Patent (cost / market / RFR / income)
- **Split founder equity** via a Slicing-Pie contribution-units model, net of ESOP
- Optionally call **Claude Code CLI** (`claude -p`) for qualitative VC-style commentary; falls back gracefully if absent
The generated xlsx is fully interactive — edit any yellow cell on the `Inputs` sheet and every downstream sheet recomputes via live Excel formulas.
## Install
```bash
source ~/projects/venv/bin/activate # or your own venv
pip install -r requirements.txt
```
## Usage
```bash
# Interactive deep evaluation
python cli.py
# JSON-driven
python cli.py --input examples/meetingroi.json --output meetingroi.xlsx
# Portfolio screening across many ideas
python cli.py --screen examples/meeting_variants_screen.json \
--screen-output portfolio.xlsx
# Dump a blank input template
python cli.py --dump-template my_idea.json
# Skip the Claude CLI call (rubric-only)
python cli.py --input my.json --no-ai
# Absorb an existing investor deck (.pptx) into JSON inputs
python cli.py --from-pptx deck.pptx --to-json deck_inputs.json
# Deterministic regex pulls money/percent anchors (TAM, revenue, IRR,
# margins, ask). If `claude` is on PATH, an LLM pass fills the rest
# (problem, solution, ICP, rubric scores). Skip the LLM with --no-ai.
# Produce an investor pitch deck (.pptx) alongside the xlsx
python cli.py --input my.json --to-pptx pitch.pptx
# 10-slide deck: Cover · Opportunity · Solution · Traction · 5-Yr
# Projections · Valuation (7 methods) · AI-Native Org · Equity ·
# Investment Ask · Closing.
# Roundtrip: absorb a competitor deck and reproduce as your own
python cli.py --from-pptx their_deck.pptx --to-pptx my_deck.pptx
```
> .ppt (legacy binary, pre-2007) is not supported directly — convert first:
> `soffice --headless --convert-to pptx your_deck.ppt` (requires LibreOffice).
## Workbook sheets
`Summary` · `Inputs` · `Scorecard` · `Org` · `Equity` · `Valuation_DCF` · `Valuation_Comps` · `Valuation_EarlyStage` · `Valuation_Patent` · `Screening` · `Definitions`
## Method choice notes
- **Modified RICE** comes from the Samyama screening template. **WSJF (SAFe)**, **CD3 (Reinertsen)**, and **EMV** are run alongside it because the four methods diverge on time-sensitive vs. dollar-payoff bets — divergence in rank is itself a signal.
- The rubric weights **AI Resilience at 17%** (heaviest). If commodity LLMs eat your wedge in 24 months, the score reflects it.
- Every role title in the org template is intentionally AI-native (e.g. `Forward-Deployed AI Engineer`, `Eval & Quality Engineer`, `Decision Intelligence Lead`).
## License
MIT — see `LICENSE`.