HDDT Framework

Hypothesis-Driven Decision Tools (HDDTs) are AI-powered systems that convert a single, testable hypothesis into a reproducible, citation-backed decision artifact through structured evidence collection, evaluation, and reasoning.

Technical Specification v1.0

Hypothesis-Driven Decision Tool (HDDT)

An HDDT transforms one decision-relevant hypothesis into a structured output by collecting evidence, evaluating support and contradictions, scoring confidence, and reporting uncertainty for action.

Motivation

  • Traditional hypothesis testing often requires weeks of manual research, expert synthesis, and high consulting cost.
  • AI lowers the cost of evidence acquisition and synthesis while keeping a transparent methodology.
  • HDDTs augment human judgment by automating evidence workflows, not replacing human decision authority.

Core Principle

  1. 1Question
  2. 2Hypothesis
  3. 3Evidence collection
  4. 4Evidence evaluation
  5. 5Support or refute
  6. 6Confidence
  7. 7Decision artifact

Methodology

Step 1

Define hypothesis

Hypotheses should be objective, measurable, falsifiable, and decision-relevant.

Step 2

Construct judgment schema

Decompose the top-level hypothesis into testable subquestions such as market, technology, competition, team, and investor fit.

Step 3

Collect evidence

Gather supporting evidence, contradictory evidence, and missing evidence from scientific and commercial sources.

Step 4

Evaluate evidence

Score each node for quality, quantity, consistency, contradictions, confidence, and unknowns.

Step 5

Score outcomes

Classify each node as support, partially support, refute, or inconclusive with a confidence interval.

Step 6

Generate decision artifact

Produce executive summary, method, evidence tables, tests, confidence, gaps, and recommendations.

Characteristics

  • Deterministic
  • Transparent
  • Reproducible
  • Versioned
  • Citation-backed
  • Investor-readable
  • Scientifically defensible

HDDT Tool Library

  • Biology Feasibility Tool
  • Financeability Tool
  • Investor-Specific Financeability Tool
  • Commercial Feasibility Tool
  • Regulatory Feasibility Tool
  • IP Feasibility Tool
  • CMC Feasibility Tool

Execution Architecture

1.Input
2.Hypothesis
3.Judgment schema generator
4.Parallel AI agents
5.MCP connectors
6.Scientific search
7.Repository and data room analysis
8.Evidence graph
9.Scoring engine
10.Decision generator
11.Markdown report
12.Notion archive
13.Website publication

The purpose of an HDDT is to improve decisions. The defensible intellectual property is the repeatable methodology, not a single report.