# Acquit.ai

> Sentient analysis for litigation intelligence. Models opposing parties as cognitive agents with behavioral modifiers to predict actions. 91% accuracy across 201 tracked predictions.

Acquit.ai is a litigation intelligence platform built by Colin McNamara (Self-Improving Code, Austin TX). It replaces sentiment analysis with sentient analysis: modeling opposing parties as cognitive agents with belief architectures, wound stacks, and behavioral modifiers to predict what they will do next.

## Core Concepts

- **Sentient Analysis**: Models opposing parties as cognitive agents with 5-tier behavioral modifier stacks, rather than just reading tone/sentiment
- **ARPN Framework**: Adapted FMEA (Failure Mode and Effects Analysis) for litigation. Extends standard RPN with Behavioral Amplifier and Cascade Reach dimensions
- **Acquit Score**: Multi-model adversarial consensus scoring. Multiple AI architectures score independently, then disagreements are investigated
- **Behavioral Modifier Framework (BMF)**: Maps 12 cognitive biases to quantified payoff adjustments in litigation game theory
- **ExhibitCTL**: Open-source (MIT) evidence collection and court-ready exhibit generation pipeline

## Key Pages

- Homepage: https://www.acquit.ai/
- About (founder story, credentials): https://www.acquit.ai/about/
- Services: https://www.acquit.ai/services/
- Get Your Acquit Score: https://www.acquit.ai/score/

## Research & Insights

- From Sentiment to Sentient (thesis): https://www.acquit.ai/insights/sentient-analysis/
- Red Team Your Case: The ARPN Framework (practitioner guide): https://www.acquit.ai/insights/arpn-framework/
- Adversarial Consensus Scoring (validation methodology): https://www.acquit.ai/insights/adversarial-consensus/
- Behavioral Modifiers in Litigation Game Theory (theory): https://www.acquit.ai/insights/behavioral-modifiers-framework/
- The Real Cost of eDiscovery (market research): https://www.acquit.ai/insights/ediscovery-cost-anatomy/

## Evidence Pipeline (6 Steps)

1. Collect - Multi-source evidence gathering
2. Verify - SHA-256 chain of custody
3. Classify - AI-powered evidence classification
4. Brief - Attorney-ready intelligence briefings
5. Present - Court-ready exhibit packaging
6. Prove - Cryptographic proof of integrity

## Open Source

- ExhibitCTL (MIT): https://github.com/colinmcnamara/exhibitctl

## Contact

- Consultation: colin@acquit.ai
- General: info@acquit.ai
- Schedule: https://calendly.com/acquit-ai/30min