Axenix BotAx

The main articles are:

• Speech Recognition (Technology, Market)
• Speech technology: On the path from recognition to understanding
• Chatbots (Chat-bot) Virtual interlocutors

2024: Bringing BotAx LLM Platform to Market

Axenix on August 14, 2024 announced the launch of the LLM platform (Large Language Model) BotAx. The generative artificial intelligence solution allows customers to use large language models to quickly build and implement chatbots.

The BotAx platform is built on a RAG architecture (Retrieval-Augmented Generation). When a certain request is received, the chatbot selects the relevant data in its own knowledge bases. Then, on this basis, creates a request for LLM, which allows you to enrich the response with additional information from third-party databases, make it more correct and complete. RAG also allows you to increase user satisfaction from the communication process by adjusting to the communication style and offering more relevant answers.

The BotAx platform gives customers the ability to create and use their own specific custom chatbots for a wide range of tasks - from HR and marketing analytics to processing regulatory documents, creating an interactive knowledge base and generating text. If necessary, Axenix specialists help correctly integrate it with the most suitable company database - relational, document-oriented, graph, etc.

There are two main use cases for the BotAx platform:

• selection and customization of a chatbot from an existing set (about ten models are already presented on the platform);
• creating your own chatbot for the tasks of each specific company.

In the future, another No-code scenario will be available - creating a chatbot from components offered on the platform without the need for programming.

"Practice shows that open source generative models trained on open Internet data contain little specific knowledge in narrow or specialized areas. They know almost nothing about a particular company. That is why RAG architecture is a solution to the problem: combining a model trained on shared data with a specific company's knowledge base and external sources. The Axenix team, assembled from experts in the field of NLP, language models, generative AI, has been working on the development of the service for about six months. We conduct test implementations, learn to work with different types of data, thus constantly improving the product, "said Vasily Krikunov, an expert in the field of Axenix AI.

According to him, another feature of the BotAx platform is that it can be used on-premium in a secure closed circuit of the enterprise with efficiency comparable to public generative models to solve specialized problems.

You have already created a consultant business assistant on BotAx; an assistant metallurgist providing information on the main technological processes and standards in steel production; chatbot to support the knowledge base on supply chains; General purpose LLM to safely solve a wide range of tasks within the company, as well as a chatbot for Self-Service BI.

"Our main task is to provide customers with a no-code chatbot designer in the future, which will allow them to correctly organize work with language models, independently select the necessary objects to create a solution for any task. Thanks to chat bots, companies can increase staff productivity, qualitatively transform the IT landscape and organizational structure, "added Vasily Krikunov.

According to him, industrial enterprises, financial organizations, retail companies and many others are showing interest in language models today.