Latent-Sense Technologies
ReX
An Advanced Reasoning Agent
Beta Release End of March 2025
At Latent-Sense Technologies, we have designed and developed neuro-symbolic AI methods to transform LLMs (any) from fluent text generators into robust analytical tools that are enabled to execute structured reasoning to detect inconsistencies, causal chains, and contradictions, within a body of knowledge (documents, corpora).
Our methods have the additional benefit of helping to guard against hallucinations. This is achieved without the need for a RAG.
Add (input) your documents and reason over your corpora (unstructured written documents) with ReX (Reasoner X), an advanced reasoning agent. ReX can be used as a standalone reasoner or integrated with your own stack to function as a reasoning agent.
We create and provide
AI building blocks
for inference and
interpretability.
Building AI Business Opportunities
Inject state of the art AI building blocks into your application or product without costly infrastructure and model training, or specialized expertise. ​Build a business model servicing your market domain and gain a competitive advantage using the latest in interpretable AI solutions. ​Interrogate your unstructured data with fast, unparalleled 'effort to insight'.
​
Achieve the benefits and ROI, at a fraction of the cost, within the same day versus weeks or months of homegrown and outsourced development.
RelsD is a lightweight, cost-effective AI building block that extracts, categorizes and computes saliency of information and relationships from vast datasets. Whether enhancing RAG systems to retrieve only the most pertinent information, empowering GenAI to ground its outputs in salient data, or enabling AI systems to make informed decisions, RelsD integrates seamlessly into existing AI stacks.
ReDid is a versatile AI-powered deidentification dual mode building block that redacts sensitive and private information, either through the common PII mode or by Intent (the user's defined concepts, ideas and information).
AiTD is a building block that detects and labels (highlights) synthetic text and can be deployed effortlessly to serve as a foundational AI technology to enhance the reliability, transparency, and effectiveness of AI-driven applications to guard against AI generated misinformation, intellectual property attribution, and compromising the integrity of data-driven systems.
Latent-Sense Technologies is a core research and development company focused on translating fundamental AI research into domain agnostic building blocks that integrate with or compliment market applications. We look to partner with academic or industry research entities as well as businesses intent on growing their AI capabilities and services.
Cutting-edge
AI building blocks
leveraging hybrid neuro-symbolic architectures and interpretive infrastructures.
Scalable
Agnostic
Domain | Model | Data
Data-Sovereign
Retain exclusive ownership of your data with total data-sovereignty.
Transparent AI
Get access to how insights are determined with explorable, and downloadable analyses.
Included with every insight!
Our cost-effective, all-inclusive AI platform, delivers 'data-to-insight' within minutes not months. With no additional or hidden costs for our customers.
Ingest & Extract
Ingest a diverse array of file types for preprocessing, indexing, segmentation and extraction. (.txt, .json, .html, .csv, with more coming soon).
Insights @ Scale
Our cognitive machines provide insights at scale across your organizations’ (un)structured text data sets.
​
Output Analysis
Download analyzed semantic spaces, visually explore saliency and relationships at scale, or ‘print’ deidentified documents in their original format and structure.
Workspace & API Console
Manage access, data, and projects from our cognitive AI platform giving you complete control of your cognitive AI pipeline including estimates and exact pricing before committing.
Compute Ready
Every cognitive insight includes the necessary underlying infrastructure (EC2, S3 etc.) and cloud services orchestration.
Deployment Paths
API
Workspace & Console
