AOI LLM for AmiVoice^® Communication Suite

Generative AI that can safely handle confidential information because it is a local environment

"AOI LLM for AmiVoice Communication Suite (hereinafter referred to as AOI LLM)" is a generative AI service that can generate and summarize call content, extract Q&A, and extract VoC (voice of the customer) in the customer's local environment without releasing any call data, including confidential information, to the outside. It is optimized for contact center operations through customization and system integration using knowledge and specialized technology based on a track record of implementation in contact centers in a wide range of industries. It can be used without restrictions such as the number of times it can be used, as with cloud-based generative AI services.

*This service isAmiVoice Communication Suite" is an optional service.

By inputting the call data converted into text by "AmiVoice Communication Suite" into "AOI LLM," it becomes possible to easily and safely generate summaries and extract Q&A, greatly improving work efficiency.

By fine-tuning the data to suit the purpose, it becomes possible to utilize the call text data in a variety of ways that suit the user's purpose. Below are some examples of how it can be used.

Generated Summary

Automatically summarize data converted into text using "AmiVoice Communication Suite."

The summary format can be freely tuned. We can summarize in the most optimal format to suit your business needs.

Q&A excerpts^*

From the voice recognition results, it is possible to extract Q&A in the form of "customer questions" and "operator responses." This makes it easy to create data that can be used to expand FAQs, etc.

VoC Extraction^*

It is possible to extract Voice of Customer (VoC) contained in phone calls. Since it is possible to flexibly tune the perspective from which to extract, such as "dissatisfaction," "gratitude," and "requests," it is now easy to extract VoC data from phone call content, which was previously difficult.

*UI support for "AmiVoice Communication Suite" is planned to be implemented in the future.

Generative AI (LLM) can be built in a local environment, allowing for secure use.

Since the generation AI (LLM) is built in a local environment, call data is never sent to the outside, so you can use it with peace of mind. You can also build a system to suit your company's environment, whether on-premise or on cloud.

By using the API of "AmiVoice Communication Suite," it is possible to smoothly integrate with a variety of functions, from voice recognition results to emotion analysis.

Generates call text data and can be used for various purposes such as summarizing, Q&A excerpts, and VoC extraction

It is possible to generate summaries in the optimal format for the customer's business operations. In addition, the results of speech recognition can be used to extract Q&A in the form of "customer questions" and "operator responses," which can be used to expand FAQs. Furthermore, it is possible to extract VoC contained within a call from various perspectives, such as "dissatisfaction," "gratitude," and "requests." This makes it easy to extract VoC data from call text data, which was previously difficult.

Fine-tune for your purpose to improve accuracy

To prevent hallucination in the answers generated by the AI, we perform fine tuning according to the purpose and create an LLM model specifically for the customer. Furthermore, we have improved the accuracy of answers by combining Retrieval-Augmented Generation (RAG), a technology that uses external information to generate answers.

You can choose the base model. You can also upgrade to the latest base model.

We apply the model that best suits your purpose from the publicly available locally available generative AI (LLM). In addition, as technology advances, you can update to the latest base model.