Advancing how we measure agent behavior

We’re building the methods and infrastructure that push agentic evaluation forward — where signals are scalable and grounded in user preferences.

Blog

Oct 7, 2025

Climbing the Hills That Matter

Why evals fall short—and how Agent Behavior Monitoring turns messy production data into the only metrics that matter.

Blog

Oct 7, 2025

Climbing the Hills That Matter

Why evals fall short—and how Agent Behavior Monitoring turns messy production data into the only metrics that matter.

Blog

Oct 7, 2025

Climbing the Hills That Matter

Why evals fall short—and how Agent Behavior Monitoring turns messy production data into the only metrics that matter.

Our Belief

You cannot improve what you cannot measure.
Continuous monitoring of user preferences and agent behavior underpins any data flywheel. This makes automated scoring a central bottleneck to self-improving AI.

You cannot improve what you cannot measure. Continuous monitoring of user preferences and agent behavior underpins any data flywheel. This makes automated scoring a central bottleneck to self-improving AI.

Read our work on “Climbing the Hills That Matter”

Read our work on “Climbing the Hills That Matter”

Judgment Labs Research

We build frontier quality measurement systems that integrate directly with our agent behavior monitoring infrastructure. Our research currently focuses on three frontiers:

Blog

COMING SOON

Building specialized scoring

Custom scorers that turn noisy data into clear signals.

Blog

COMING SOON

Building specialized scoring

Custom scorers that turn noisy data into clear signals.

Blog

COMING SOON

Building specialized scoring

Custom scorers that turn noisy data into clear signals.

Blog

COMING SOON

Discovering signals that matter

Uncovering user preferences hidden in real trajectories

Blog

COMING SOON

Discovering signals that matter

Uncovering user preferences hidden in real trajectories

Blog

COMING SOON

Discovering signals that matter

Uncovering user preferences hidden in real trajectories

Blog

COMING SOON

Advancing how we measure behavior

Aligning judges to align with expert review and user preferences

Blog

COMING SOON

Advancing how we measure behavior

Aligning judges to align with expert review and user preferences

Blog

COMING SOON

Advancing how we measure behavior

Aligning judges to align with expert review and user preferences

Applied Research Success Stories

Case Study

Oct 1, 2025

Improving Legal AI with Judgment

Boosting legal drafting quality and speeding agent releases with Judgment.

Case Study

Oct 1, 2025

Improving Legal AI with Judgment

Boosting legal drafting quality and speeding agent releases with Judgment.

Case Study

Oct 1, 2025

Improving Legal AI with Judgment

Boosting legal drafting quality and speeding agent releases with Judgment.

Case Study

Sep 28, 2025

Transforming Internal Agents with AI

Cutting trajectory bloat and boosting agent efficiency with Judgment.

Case Study

Sep 28, 2025

Transforming Internal Agents with AI

Cutting trajectory bloat and boosting agent efficiency with Judgment.

Case Study

Sep 28, 2025

Transforming Internal Agents with AI

Cutting trajectory bloat and boosting agent efficiency with Judgment.

Case Study

Sep 25, 2025

Real Time Monitoring of Financial AI

Restoring trust and accuracy in financial AI with real-time monitoring.

Case Study

Sep 25, 2025

Real Time Monitoring of Financial AI

Restoring trust and accuracy in financial AI with real-time monitoring.

Case Study

Sep 25, 2025

Real Time Monitoring of Financial AI

Restoring trust and accuracy in financial AI with real-time monitoring.

Other Works

Publication

May 7, 2025

Osiris: A Lightweight Open-Source Hallucination Detection System

Alex Shan, John Bauer, Chris Manning

Publication

May 7, 2025

Osiris: A Lightweight Open-Source Hallucination Detection System

Alex Shan, John Bauer, Chris Manning

Publication

May 7, 2025

Osiris: A Lightweight Open-Source Hallucination Detection System

Alex Shan, John Bauer, Chris Manning

Blog

COMING SOON

A Framework for Post-Building

A novel approach to agent post-building.

Blog

COMING SOON

A Framework for Post-Building

A novel approach to agent post-building.

Blog

COMING SOON

A Framework for Post-Building

A novel approach to agent post-building.

Blog

COMING SOON

An LLM Judge is All You Need

Exploring reward modeling in non-verifiable domains with LLM judges

Blog

COMING SOON

An LLM Judge is All You Need

Exploring reward modeling in non-verifiable domains with LLM judges

Blog

COMING SOON

An LLM Judge is All You Need

Exploring reward modeling in non-verifiable domains with LLM judges

© 2025 Judgment, Inc.

Resources

Docs

Company

Home

Research

Pricing

Contact

Community

GitHub

LinkedIn

X

© 2025 Judgment, Inc.

Resources

Docs

Company

Home

Research

Pricing

Contact

Community

GitHub

LinkedIn

X

© 2025 Judgment, Inc.

Resources

Docs

Company

Home

Research

Pricing

Contact

Community

GitHub

LinkedIn

X