Sentius

Sentius is an agentic automation platform for enterprises that lets teams automate individual tasks and entire workflows using autonomous AI agents operating under human supervision. Its core product line, announced as the Sentius Teach & Repeat Platform, includes a Browser Agent that executes natural-language commands by controlling a Chromium-based browser, an OpenAPI Agent that calls existing systems over REST APIs, and a Prompt Agent that routes intelligence tasks to large language models (with MCP tools). Desktop and Document agents are described as under development. Around the agents, Sentius Studio provides an application for designing, testing, editing, deploying and monitoring automations in test and production environments, and a Workflow Engine runs those automations at organizational scale, including inside a customer's secure perimeter.

The company frames its approach around "UI maps" used as blueprints for agent execution, which it says yields fewer failures, lower token costs and full audit trails compared with unguided agent reasoning. It positions the platform against two limits of prior automation: rule-based robotic process automation that breaks as web applications change, and the absence of any automation path for previously unseen websites. The Browser Agent is described as combining rule-based and neural approaches so it can learn a task by imitation from a demonstrated sequence of steps or work out an unknown task on its own, then save successful executions as reusable instructions.

Marketed uses span automation and execution (for example creating invoices in QuickBooks using customer data from Salesforce, populating HR forms, synchronizing ERP and CRM records), testing and quality (regression checks in SAP or Oracle, enforcement of approval steps, WCAG/ADA accessibility validation, detection of broken UI elements), and knowledge capture and onboarding. Named use cases include compliance screening in financial services, preparation of client proposals by relationship managers, and accounts-receivable forecasting in treasury management.

Founding story

Sentius was founded by Daniel Kornev, who serves as CEO and who is described as having managed 20 teams that shipped a personal AI assistant with 60 million daily active users, and Dr. Mike Burtsev, Chief Science Officer, an AI Fellow at LIMS, founder of DeepPavlov.ai and author of more than 100 machine learning and AI papers. The company name is drawn from the Latin "sentio" (to feel, perceive or think).

Business model

Sentius sells an enterprise agentic automation platform. Components include per-agent capabilities (Browser Agent, OpenAPI Agent, Prompt Agent), the Sentius Copilot+ desktop application for Windows, macOS and Ubuntu, the Sentius Studio authoring and monitoring application, and a Workflow Engine service that executes automations at organizational scale; the agents and the workflow engine are also reachable through APIs. Access is currently offered through a private beta and an early-access ('EA') program, with deployment inside a customer's secure perimeter, on employee devices or in virtual desktop infrastructure.

The sources describe an enterprise platform delivered as applications plus an API-accessible workflow engine service, offered via private beta and an early-access program; pricing terms are not stated.

Traction

A private beta of the Teach & Repeat Platform was reported in December 2024 as already in use by design partners and customers. The company states that open source AI products its team has worked on total more than 722,000 installs, over 6,700 GitHub stars, more than 105,000 NLP model downloads and over 1 million DockerHub pulls, and that team research has appeared at ACL, NeurIPS and EMNLP.