How accurate is the facial recognition?

Agent-R's Vision AI detects and tracks each face in real time even when several people approach simultaneously, and it recognizes returning visitors to offer a personalized greeting. It operates reliably across varied lighting conditions and multi-directional approaches.

2026 MSIT-recognized AI supplier

NIPA-registered AI startup

The AI reception agent that
sees, listens, and responds like a person

Name: neurosam.AI Agent-R
Brand: neurosam.AI
Availability: InStock

24/7 reception with zero added headcount — one platform that sees, hears, and responds

neurosam.AI Agent-R is an all-in-one hybrid SaaS solution that unifies an AI reception agent with 24/7 AI phone answering. By combining facial recognition, voice AI, and telephony AI, it delivers unmanned guidance and intelligent customer reception across smart offices, hotels, hospitals, and stores — every space that welcomes people.

Request a demo Explore the solution

Reply within 24 hours · Free technical materials · No obligation

LIVE · 24/7 unmanned reception

An AI that talks with you by voice

AI-generated concept image · deployment environment visualization

The reception burden on the front line

Three realities facing businesses, stores, and institutions
that must run 24/7 without reception staff

You're short on staff for nights and holidays

Early-morning check-ins, inquiries on clinic closed days, reservation calls during off-hours — you can't grow labor costs without limit, so reception gaps keep piling up.

See hotel, retail & healthcare cases

Kiosk accessibility mandates are imminent

Under the amended Korea Disability Discrimination Act, starting 2026-01-28 an existing 450,000 units must meet voice, multilingual, and accessibility requirements.

See the government & public-sector approach

Multilingual reception staff are hard to find

Hotels, tourism, and public service desks must handle Korean, English, Chinese, and Japanese around the clock, yet hiring more people is a heavy cost burden.

See multilingual reception cases

Agent-R Platform

A 24/7 unmanned reception AI kiosk agent robot
— it runs as a domain persona

Agent-R is a 24/7 unmanned reception solution that unifies on-premise service robots and AI callbots on a single platform. Swap only the site-specific system prompt and RAG knowledge base, and it instantly runs as a new domain persona — so AI covers every customer touchpoint, from greeting visitors to handling missed calls.

AI facial recognition & vision perception

Even when several people approach at once, it instantly recognizes each face and greets returning visitors by name with a personalized welcome. When someone loses interest and turns away, it naturally wraps up the conversation — reception with autonomous judgment.

Multi-Face Detection Returning-visitor recognition Autonomous engagement

Seamless voice conversation

Even in the brief moment while the AI prepares a response, it keeps the conversation flowing with natural filler phrases. The instant a user starts speaking, it switches to listening mode — so the exchange feels just like talking with a person.

Nudge prompts Cut-in support 480ms response

24/7 AI phone assistant & AI callbot

The AI phone assistant answers calls placed to your company's main line on your behalf. It responds accurately using a customer-specific knowledge base (RAG), and after each call it automatically generates a conversation summary and follow-up actions for the right team member. It runs nonstop through nights and holidays.

AI phone answering RAG-based answers Auto summary/handoff

Unified monitoring dashboard

Manage the reception status of tens to hundreds of robots and phone channels at a glance. It supports permission separation between headquarters and branches, and you can receive real-time SMS/email alerts for important reservations.

Multi-Tenant Role-based permissions Real-time alerts

Domain Persona Engine

A different persona for every site,
the same platform.

Conventional kiosks and chatbots had to be rebuilt from scratch for every new site. Agent-R decouples the system prompt and RAG knowledge base into modules, so you only define the site-specific persona and the same platform deploys quickly to a new site. This makes the SaaS economics of building once and scaling to many sites genuinely possible.

Modular personas

The system prompt defines personality, tone, and role, while the RAG knowledge base injects domain knowledge separately.

Fast new-domain deployment

When adopting a new site, swapping the persona alone lets you apply to a new domain in a short timeframe.

Persona library

Domain personas across healthcare, hotels, public sector, retail, and more accumulate and expand into a marketplace.

Customer data isolation

Per-site RAG index separation provides data isolation and access control as a default design.

Core Technology

neurosam.AI core technology capabilities

We focus on the software technology that builds the "brain" of robots and AI systems.

AI Software Architecture

We design robust backends for intelligent autonomous systems and build highly scalable architectures that adapt flexibly to a wide range of hardware.

Domain Persona Engine

On top of a multimodal pipeline integrating Vision AI, STT, TTS, and LLM, we decouple domain personas into modules so the same platform operates with a different personality at every site.

Strategic AI Advisory

We provide partners with optimized AI solutions and deliver expert advisory on service robot adoption and interaction technology.

Why neurosam.AI

A category where law, market, technology,
and trust are all aligned

The unmanned reception market is at the exact moment when legal mandates and government policy momentum flow directly into this business. neurosam.AI has already cleared certification as a government-recognized AI supplier twice, and is rapidly accumulating its own IP and open-source assets.

A category where law & market momentum are aligned

Under the amended Korea Disability Discrimination Act, starting 2026-01-28 about 450,000 existing kiosks in Korea must meet voice, multilingual, and accessibility requirements. Agent-R is virtually the only category that inherently satisfies voice conversation, multilingual guidance, and accessibility for the visually impaired.

Accessibility mandate 450,000 kiosks Inherently accessible

Government-recognized AI supplier

Recognized by Korea's Ministry of Science and ICT (MSIT) and NIPA as a certified AI supplier in 2026 — an external validation of the company's technology and delivery capability that passed government screening within months of incorporation.

MSIT-recognized NIPA-registered Government-screened

IP assets in motion + open source released

About 1.5 months after incorporation, we completed simultaneous trademark filings across 4 classes (09 · 35 · 41 · 42), putting brand protection across the entire business scope in motion. At the same time, we open-sourced Tow, our Go-based operations infrastructure backbone, securing an external path for technical validation.

Trademarks in 4 classes Open-source Tow Tech backbone open-sourced

Founder & senior engineer + a proven R&D track record

A KakaoBank alumnus, our founder is an entrepreneur and senior engineer with two co-founded ventures and multiple CTO roles and a track record of 5 government R&D projects delivered successfully (3 consecutive years of TIPS + 2 Global Market-Type R&D projects). Full career details are available in the About section.

3 consecutive years of TIPS 5 government R&D projects Two co-founded ventures

Deploy a new site persona in about a week — ready when you are

Book a consultation

About Us

Connecting LLMs to
the physical real world

neurosam.AI is a software-first AI brand operated by Neurosam AI Inc. From digital agents to physical robots, we engineer the core intelligent systems that let AI perceive on its own and make autonomous decisions.

A software-first approach for flexible, scalable AI systems

Optimized service robot and kiosk interaction

2026 MSIT-recognized AI supplier — government-screened technology credential

NIPA-registered AI startup — government certification as a youth AI startup

Trademark filings in 4 classes + open-source Tow released — IP & technology assets in motion

Led by a founder & senior engineer — KakaoBank alumnus + two co-founded ventures + 5 government R&D projects delivered successfully

Founder & CEO

Murry Jeong · 정문창

Dual bachelor's degrees in Industrial Engineering and Computer Engineering from Ajou University. A KakaoBank alumnus, entrepreneur, and senior engineer with two co-founded ventures and multiple CTO roles, he currently leads R&D on a domain-persona-based robot assistant that integrates Vision AI, STT, TTS, and LLM.

Government R&D track record

5 government R&D projects delivered successfully — 3 consecutive years of TIPS Startup Growth Technology Development (2022 · 2023 · 2024) + 2 years of Global Market-Type Startup Technology Development (2014 · 2015)

Government & industry-academia governance

Member, Curriculum Innovation Committee of Ajou University's SW-Centered University · Industry-Academia Cooperation Fellow (2022–) · SW Maestro mentor · FastCampus MLOps instructor

Technology assets

Open-source Tow released (Mar 2026) · trademark filings in 4 classes (Feb 2026) · the neurosam.AI engineering blog published regularly

LinkedIn Blog Tow

Leave your email and we'll send you the technical materials

Product one-pager · adoption process guide · environment assessment. No obligation to adopt.

Reply within 24 hours · Free technical materials · No obligation · Request a 1:1 consultation

FAQ

Frequently asked questions

What is the difference between an AI reception agent and a regular kiosk?

A regular kiosk is a static, touchscreen-based interface. Agent-R, by contrast, proactively perceives visitors through facial recognition, responds with natural-language voice conversation, and judges interest to autonomously start or end a conversation. It delivers intelligent unmanned guidance that sees, listens, and reasons like a person.

How does 24/7 AI phone answering work?

The AI phone assistant answers calls placed to your company's main line on your behalf. It responds accurately using a customer-specific knowledge base (RAG), and after each call it automatically generates a conversation summary and follow-up actions and delivers them to the right team member. It runs without interruption at night, on holidays, and whenever staff are away.

Which industries can use Agent-R?

It can be used in every space that welcomes people — smart offices (AI visitor management, staff paging), hotels & resorts (24/7 unmanned concierge, unmanned check-in), retail & F&B (in-store guidance, voice ordering), and healthcare institutions (unmanned intake, wayfinding to exam rooms).

What level is the facial-recognition-based visitor management?

Agent-R's Vision AI detects and tracks each face in real time even when several people approach simultaneously, and it recognizes returning visitors to greet them by name with a personalized welcome. It enables AI facial recognition access control that operates reliably across varied lighting conditions and multi-directional approaches.

How long does deployment take?

Agent-R's key advantage is its domain persona engine. Swap only the site-specific system prompt and RAG knowledge base, and the same platform runs as a new site persona — so persona deployment is possible within a week for a standard adoption scenario. The full timeline, including hardware installation and system integration, depends on your environment and is shared during the consultation stage.