SAN JOSE, Calif., Sept. 22, 2026 (GLOBE NEWSWIRE) — Kognitos, the pioneer of neurosymbolic AI for business and finance automation, today announced a partnership with TypeSafe AI to make its Jev decision model accessible to business users. Teams can describe their requirements in English, apply Jev’s specialized decision intelligence to the task, and govern how those decisions are used within auditable business processes.

The integration is already being deployed for ticket classification, insurance claim workflows, and AR collections inbox sorting. It extends a platform that many enterprises now use for agentic AI automation across finance, supply chain, HR, and operations. Connecting a language model to a workflow leaves teams responsible for interpreting its outputs, handling uncertainty, and controlling what happens next. Kognitos and TypeSafe bring those responsibilities into an executable process that business users can understand and govern in English.

Kognitos’s English-as-Code builder, Quill, selects the appropriate tools for each task and composes them into an executable business process. English as Code provides the overarching governance framework: defining how those tools operate together, how their outputs drive actions, and when human judgment is required.

“Business users should be able to describe the work in English and have the right intelligence applied to each task,” said Binny Gill, founder and CEO of Kognitos. “Quill does the hard work of selecting and combining the tools. Jev brings structured decisions with calibrated confidence, and English as Code governs how those decisions become actions. That is how we put specialized intelligence in the hands of business users while keeping them in control. Governance is built into the architecture.”

The right intelligence for each business task

Enterprise processes rarely depend on one kind of intelligence. They may need an AI model to understand language or make a judgment under uncertainty, then deterministic logic to handle calculations, apply business rules and decide what happens next. Kognitos brings those different methods into a single governed process.

During process building, Quill selects and composes LLMs for language understanding and generative reasoning, Jev for structured probabilistic judgments, deterministic mathematics for calculations, and workflow capabilities for business rules and actions. This approach is designed to improve accuracy and reduce cost by matching each task to the appropriate computational method.

Jev returns typed decisions with calibrated probabilities and confidence scores. Kognitos makes those capabilities accessible through English and incorporates their outputs into explicit business rules. Teams can define confidence thresholds for automated action, require human review when confidence is insufficient, and trace decisions and resulting actions through the same audit trail.

In an AR collections inbox, for example, Jev can classify an incoming email as a payment commitment, a dispute, or a request for documentation. English-defined rules then govern routing and follow-up, including escalation when the required confidence threshold is not met.

That division of labor is central to the integration. Jev makes the probabilistic judgment, while Kognitos’s deterministic English-as-Code framework governs what happens next. Business users can read, review, and refine the governing logic without working with application code or model APIs.

About TypeSafe AI

TypeSafe AI develops System One Models for decisions inside software. Its first model, Jev, transforms unstructured inputs into structured probabilistic decisions with calibrated confidence, supporting classification, routing, scoring, and other decisions under uncertainty. Learn more at typesafe.ai.

About Kognitos

Kognitos automates business operations with the first neurosymbolic AI platform engineered for deterministic, hallucination-free execution at enterprise scale. Built for finance, supply chain, and operations teams that cannot tolerate probabilistic errors, Kognitos turns tribal and system knowledge into documented, AI-refined automations using English as code, creating a dynamic system of record that enhances productivity and decision-making. Every workflow is readable, auditable, and governed in English, with humans in control of the logic that runs the business. With its patented Process Refinement Engine, Kognitos delivers faster ROI, lower costs, and empowered teams across hundreds of enterprise use cases. Headquartered in San Jose, California, Kognitos is backed by leading investors including Khosla Ventures, Prosperity7 Ventures, Engineering Capital, and Wipro Ventures.

Media Contact:
pr@kognitos.com


Primary Logo

About The Author