AUTONOMOUS INTELLIGENCE is a working brand name. We are preparing incorporation in Jeju, Korea. About the companyPre-incorporation
AUTONOMOUSINTELLIGENCE
PROJECTS

Five projects, five different stages

Every project states its development stage. Prototypes are internal validation work and do not imply a commercial product or certified performance.

Working PrototypeResearch & Design

Prototype performance figures are not published yet. Once methods and conditions are documented, they will be published per project.

01

AI Browser & Computer Use

A browser agent that navigates and operates web work with no API

Working Prototype

Project Overview

An agent for web interfaces without public APIs — booking channels, admin pages, public portals. Given a goal, it navigates, enters data and confirms the result.

Problem

Much of the repetitive web work in hospitality is still done by hand because there is no API. Channel management, booking checks and settlement downloads sit on screens that change often, which makes simple macros fragile.

Technical Approach

A headless browser identifies elements using both the accessibility tree and screen capture, and verifies screen state after each step. On failure it records the cause and escalates for approval rather than retrying blindly.

Key Features

  • Goal-driven navigation and multi-step forms
  • Login, session and re-authentication handling
  • Screen-state based execution checks
  • Execution logs and screenshots

Development Status

Working Prototype — behavior validated on internal workflows. Resilience to site changes and exception handling are still incomplete.

Future Development

Executor hardening, per-site adapters, finer human-in-the-loop points, auditable execution reports.

02

AI Hospitality Management

A hospitality operations agent tying bookings, rates, guest replies and settlement into one flow

Working Prototype

Project Overview

A hospitality operations system that gathers bookings from multiple channels and handles rates, inventory, guest replies and settlement in one flow.

Problem

Lodging operations split booking checks, rate changes, guest replies and settlement reconciliation across tools, so the same information is entered several times. In peak season a missed check becomes lost revenue and a guest complaint.

Technical Approach

Channel data is normalized around the PMS/CMS, and booking events trigger rate, inventory and reply tasks. Guest replies are drafted within policy; refunds and compensation require human approval.

Key Features

  • Channel booking collection with duplicate and gap detection
  • Rate and inventory adjustment workflows
  • Inquiry classification and reply drafts
  • Settlement reconciliation and variance reports

Development Status

Working Prototype — trialed on selected tasks at our own operated property. This is not a commercial service offered to other properties.

Future Development

Broader OTA integrations, stronger rate rules, multi-property support, more accurate automatic reconciliation.

03

Public Data Infrastructure

A data layer that collects, normalizes and verifies public APIs until they are usable

Working Prototype

Project Overview

Infrastructure that reliably collects public APIs and datasets, unifies schemas that differ by agency, and makes the result usable directly in business systems.

Problem

Public data differs by provider in refresh cycle, field definitions and authentication, so it is hard to use as-is. When collection code is scattered, data quality cannot be guaranteed.

Technical Approach

Collector, normalizer and verifier are separated, and every record carries source, collection time and version. Records failing quality rules are quarantined, and schema changes raise an alert.

Key Features

  • Per-agency collectors and schema mapping
  • Required-field, format and duplicate checks
  • Change detection and refresh history
  • Query API for business systems

Development Status

Working Prototype — collection, normalization and verification implemented in internal pipelines. Not operated as a public API service.

Future Development

More connectors, data lineage visualization, quality dashboards, and an external service model to review.

04

Enterprise AI & ERP Automation

Back-office automation connecting ERP and document systems

Working Prototype

Project Overview

A project that reconciles data scattered across ERP, accounting and document systems, and lets agents handle repetitive entry, settlement and reporting work.

Problem

Back-office work spends most of its time moving and reconciling data between systems. Even with automation, an exception sends a person back to the beginning.

Technical Approach

Work is split into read, verify and write stages, with human confirmation of reconciliation before any write. Every change is written to an audit log so it can be reversed.

Key Features

  • Cross-system reconciliation and variance reports
  • Settlement and invoice checking assistance
  • Document generation and form filling
  • Approval-gated writes with audit trail

Development Status

Working Prototype — reconciliation and document automation validated in limited scenarios. No client deployments.

Future Development

Standard ERP connectors, a refined permission model, an exception-handling library, on-premise deployment options.

05

MANPASIK Distributed Intelligence

Independent research into distributed intelligence where agents collaborate and cross-verify

Research & Design

Project Overview

MANPASIK explores a distributed structure in which several agents divide roles and verify each other's results, instead of scaling a single model.

Problem

In a single-agent structure, long tasks inflate context and make self-detection of errors difficult. Cost and latency also concentrate at one point.

Technical Approach

Planning, execution, verification and memory are distributed across distinct agents with consensus and falsification procedures. Work distribution to local and edge inference is examined through simulation.

Key Features

  • Role-based division of agent labor
  • Cross-verification and falsification protocol
  • Work distribution and routing hypotheses
  • Edge inference scenarios

Development Status

Research & Design — idea and architecture design. Not a finished product, validated performance, or commercial technology.

Future Development

Publish design documents, summarize simulation results, review a prototype in a constrained environment, research collaboration.

Looking for partners to validate with

If you have a real problem in hospitality operations, public data or enterprise back-office automation, we can validate it together through a pilot or technical partnership.