A first in the contact center industry!
Added a new feature that allows multiple AI agents to assist operators in real time.

On May 27th, we released "AmiVoice Communication Suite 4.6," adding a new "AI Assistant function."
The "AI Assistant function" is an industry-first feature* in which multiple AI agents perform optimal inferences tailored to the content of the call and assist operators in real time by performing tasks such as launching applications as needed.

In the contact center industry, while the use of generative AI is advancing, the increasing burden on operators due to the need to adapt to new systems and the diversification and sophistication of workflows is becoming a challenge. Furthermore, the scope of generative AI use is often limited to post-call summarization, and it is not being fully utilized to optimize the entire business process, including in-call support and subsequent tasks.

In this context, we have added an "AI Assistant function" to "AmiVoice Communication Suite" that allows operators to receive real-time support from AI agents without having to perform any special operations, based on the content of the conversation during a call. Multiple AI agents can provide the necessary information and support content at the optimal time, and support operations by performing tasks such as launching applications as needed, so operators can concentrate on their core job of interacting with customers without being strongly aware of the presence of the generated AI.

■ Features of the "AI Assistant Function"

1. Real-time support from AI agents enables an environment where employees can focus on their core business of customer service.
At the optimal moment during a call, the conversation between the customer and the operator is relayed to the AI ​​agent. Based on the content of the conversation, the AI ​​agent suggests and executes necessary processes such as summarizing, FAQ search, and information extraction. This allows operators to concentrate on their core task of answering calls without being overly aware of the AI's presence, and enables faster and more accurate responses.

2. The timing of when the AI ​​agent begins providing assistance can be flexibly set.
The AI ​​agent's timing for offering assistance can be flexibly set using five triggers: "start call," "end call," "end call," "start hold," and "keyword." For example, if you speak a pre-registered keyword such as "I'll look into it," a notification such as "[Answer candidates have been generated]" or "[Displaying FAQs]" will appear in a pop-up window on your PC. You can also adjust the display time of the notifications and switch the AI ​​agent's processing between manual and automatic.

3. Utilizing "AI multi-stage inference function" to handle complex processing.
Combined with the "AI Multi-Stage Inference Function" launched in November 2025, it is possible to perform complex processing utilizing multiple generative AIs. We can provide various real-time support tailored to your requirements.
*Click here for details on the "AI Multi-Stage Inference Function"
https://www.advanced-media.co.jp/newsrelease/10910/

4. Streamlining subsequent operations through integration with external systems.
The AI-generated data extracted during a call can be integrated with external systems such as CRM (Customer Relationship Management) systems. This automates data registration and update processes, improving accuracy and efficiency not only during calls but also in subsequent tasks.

■ Examples of how to use the "AI Assistant function"

1. Link keywords from conversation content to the FAQ system to provide the most relevant knowledge.
At any time the operator chooses, the conversation content up to that point can be shared with the AI ​​agent, which automatically extracts the most suitable keywords to search for necessary knowledge and links them to the FAQ system, enabling smooth information provision.

2. Determine the call reason and organize and optimize the necessary verification items.
By integrating with the "AI multi-stage inference function," it is possible to present the necessary confirmation items for each call reason. When an operator utters a specific keyword such as "I will confirm," the call content up to that point is transmitted to the generating AI. The first-stage generating AI analyzes the call content and identifies requirements such as "issuance of balance certificate." Based on the identified requirements, the second-stage generating AI can organize the acquired information and present any missing items.

3. Reduce the amount of data entry into the CRM by integrating information extracted during calls.
The AI-generated data, which extracts information such as addresses and contract details from call content, can be automatically reflected into external systems such as CRM (Customer Relationship Management) systems using RPA (Robotic Process Automation). This reduces the time spent on data entry and post-call processing.

*According to our research. As of the end of May 2026. "A function in contact center solutions where multiple AI agents provide support at the optimal timing depending on the situation during a call."