U.S. 12,008,332 B1 patent record (opens in a new tab)
Who we are
Scientific analysis of AI answers.
Modulus combines rigorous research, technical investigation, and hands-on problem management to address inaccurate AI answers, misleading recommendations, and unjustified exclusion affecting people, companies, and products.
Our approach
Our work begins with the business consequence. We examine why an answer is wrong, why a recommendation may be unjustified, or why a qualified business is being left out. Then we develop and carry through the response with the people responsible for the organization.
Heritage
An engineering heritage dating to 1997
The Modulus group of companies, founded in 1997, has served clients in more than 100 countries. Its work spans critical infrastructure, natural language processing, artificial intelligence, and high-performance computing. Modulus AI, Inc. applies that technical foundation to consequential AI failures.[1][2]
- Countries where Modulus has served clients
- 100+
- Critical infrastructure
- Natural language processing
- Artificial intelligence
- High-performance computing
Research
Patented LLM research
Our technical foundation includes issued U.S. patents on controllable summarization using large language models. That research informs an evidence-led practice focused on how AI systems interpret information and produce consequential answers.[3][4]
Selected issued patents
U.S. 12,380,287 B2 patent record (opens in a new tab)
Systems for controllable summarization of content
Engagement
Senior involvement throughout the engagement
Each client works directly with the professionals responsible for the investigation. Weekly working sessions connect the research to business priorities, approved actions, and the practical realities of implementation.
Our process
Immediate results that build
Modulus combines scientific research and legal support to understand and improve how AI represents you, delivering immediate results that build over time.
Although providers such as OpenAI, Google, and Anthropic are the only ones that can directly modify their proprietary model weights, Modulus uses documented evidence and scientific analysis to support formal emergency correction requests, addressing problems that source changes alone cannot resolve.
Every engagement includes weekly consultations with senior researchers. We offer enterprise services and an affordable SMB tier for small and midsized businesses and independent professionals.
Evidence and counsel
Technical evidence for counsel
Where a matter requires formal action, Modulus works with experienced counsel and supports the response with documented technical evidence. In litigation and other proceedings, Modulus experts provide independent technical reports and testimony.
Leadership
Richard Gardner
Founder and CEO
Richard Gardner founded Modulus in 1997 and has spent three decades developing complex software and high-performance systems. His work spans artificial intelligence, systems architecture, and mission-critical technology. He is a named inventor on patented large-language-model technology.[1][3][4]
At Modulus, he directs the firm’s technical approach to investigating consequential AI failures, evaluating the evidence, and developing corrective strategies with clients and their advisers.
Recognition
Three decades of recognized expertise
Modulus and its leadership have been recognized as subject matter experts in AI and financial software development for three decades. Media and industry coverage has included Bloomberg, Reuters, Forbes, International Financial Law Review, and other regional and industry publications. Most industry coverage includes High Performance Computing and AI. Modulus has also provided expert editorial review for technology books published by Wiley & Sons.
Modulus has worked in artificial intelligence and natural language processing since 1997. In 2009, the firm developed a sentiment analysis platform.[5] In 2013, Modulus partnered with IBM Watson to develop SharpeMind.[6]
Sources
- [1]Modulus • Company history (opens in a new tab)Company-reported history
- [2]Modulus • LLM development (opens in a new tab)AI and high-performance computing practice
- [3]U.S. Patent 12,008,332 B1 • Systems for controllable summarization of content (opens in a new tab)Issued June 11, 2024
- [4]U.S. Patent 12,380,287 B2 • Systems for controllable summarization of content (opens in a new tab)Issued August 5, 2025
- [5]Modulus • Sentiment analysis (opens in a new tab)Built 2009
- [6]Modulus • SharpeMind (opens in a new tab)Built in partnership with IBM Watson • 2013