Analyzes and controls the context, emotions, and goals of a conversation in real time. Introducing a next-generation AI agent that delivers next-generation customer touchpoints through “AI Call Center (phone),” “AI Chatbot (web),” and “AI Avatar (web).”
TIGEREYE Inc. (Headquarters: Chuo-ku, Tokyo; CEO: Gaku Kamimura) announces that in May 2025 it officially released “TIGEREYE AGENT,” an AI agent that understands facial expressions, voice, and text holistically to enable natural conversation.
What is TIGEREYE AGENT?
TIGEREYE AGENT is a business support AI that covers three channels—call centers, web avatars, and chatbots—and makes customer service more automated, more accurate, and more efficient at the same time.
“TIGEREYE AGENT” provides the following three solutions in a single package.
1. AI Call Center (phone)
Combining automatic speech recognition (ASR), text-to-speech (TTS), and large language models (LLMs), it understands, responds to, and records calls in real time. AI summarizes and scores call content and integrates with CRM and other systems to automate customer service.
2. AI Avatar (web)
2D, 3D, and photorealistic AI avatars that combine facial expressions, voice, and gestures to offer an experience on the web that feels like “talking to a real person.” Ideal for customer service, education, exhibitions, and other situations that call for realistic conversation.
3. AI Chatbot (web)
A high-performance chatbot that can be easily embedded in websites and internal systems. It uses each company’s own FAQs and manuals through RAG to provide instant, accurate answers. Scenarios can be designed in a GUI, and workflows can be automated without code.
Learn more about TIGEREYE AGENT
The product is equipped with a proprietary LLM architecture based on TIGEREYE’s patent-pending “context-adaptive AI technology,” which scores users’ comprehension, interest, and emotional state in real time and dynamically optimizes responses, delivering an unprecedented conversational experience.
Key feature
One of the key features of TIGEREYE AGENT in this release is its “ability to accurately sort calls through natural AI conversation alone.”
Traditionally, when handling phone and chat inquiries, human operators listened to the user’s request and then categorized it and transferred it to the department in charge.
TIGEREYE AGENT fully automates this process and handles conversations adaptively, as follows:
- Automatically recognizes the intent of a call, such as “application / inquiry / consultation / complaint / request,” from what is said
- Dynamically controls the conversation flow, such as “re-explanation, closing, or an alternative proposal,” based on the speaker’s emotions and comprehension
- Automatically routes to the appropriate handling path through FAQ, knowledge base, and CRM integration
- Automatically assigns metadata such as “satisfaction score,” “topic tags,” and “handling difficulty” to call content
- Provides a consistent sorting experience across phone, chat, and avatar channels
This addresses labor shortages, standardizes response quality, and improves user satisfaction at the same time. Adoption is growing especially in municipalities, healthcare, finance, and retail, delivering a conversational experience that “keeps no one waiting and leaves no one confused.”
Patent-pending technology: “Context-adaptive AI system”
TIGEREYE AGENT is equipped with the following patent-pending core technologies.
Trinity Eye: a role-based LLM framework
- Talk-LLM (Orbis): Generates conversational text
- Judge-LLM (Veritas): Scores the comprehension, reliability, and emotion in the conversation
- Talkflow-LLM (Sententia): Automatically controls how the conversation progresses according to its goal
Multi-layer scoring and feedback loop
- Estimates comprehension, interest, and emotional intensity from facial expressions, voice, and what is said, and turns them into quantitative scores (e.g., 0.83)
- Decides the next dialogue stage (re-explanation, closing, alternative proposal) based on real-time evaluation of responses
These functions are patent-pending as TIGEREYE’s proprietary “TALKFLOW control algorithm” and form the foundational technology for intelligent control of the dialogue engine.
Flexible choice of LLMs
TIGEREYE AGENT can use not only internet-connected, API-based LLMs but also local LLMs and fine-tuned in-house models. Because the optimal configuration can be freely chosen according to security requirements, network architecture, response speed, and other factors, it can be deployed across a wide range of industries and environments.
Expected use cases
| Industry | Use cases | Key benefits |
|---|---|---|
| Healthcare | Reception, intake questionnaire support, patient guidance | Automated responses, higher satisfaction |
| Municipalities and finance | Guidance on applications, contract support, FAQ handling | Addressing labor shortages, consistent service |
| Retail and e-commerce | Product introductions, stock information, cross-sell proposals | Better customer experience, higher sales |
| Education and HR | Interview support, e-learning support | Operational efficiency, personalization |
| Tourism and hospitality | Reservations and multilingual guidance, facility support | Higher service levels, lighter staff workload |
Flexible system configuration
- TigerStream Engine: Real-time responses through high-speed token control and asynchronous processing
- RAG Ready: Dynamic integration with business databases and internal knowledge bases
- READY series: Also includes various support functions such as YOMIGOE Ready (optimized text-to-speech readout) and Persona Ready (character management)
Comment from Gaku Kamimura, CEO and CEO
“We are working toward a world where ‘AI understands and guides people.’ TIGEREYE AGENT is the first step toward that goal. With our patent-pending technology, we will bring a new experience to the workplace, in which AI ‘reads the room and speaks with purpose.’”
Start by telling us about the challenges you want to solve with AI
Toward workplaces where AI is simply part of everyday work.