We are pleased to announce that Allganize has raised $3.4M in our series A funding round led by SparkLabs Ventures, Global Brain, along with Bass investment, Laguna Investment and Fast Investment.

Allganize provides natural language understanding AI with the best-in-class precision of Named Entity Recognition, Intent Classification and conversational AI chatbot.

Our customers, KDDI, Persol and many more, witnessed our high precision Named Entity Recognition and Intent Classification which enable our chatbot to understand and process complex natural language as if it was human agent.

Allganize's focus is and always will be Natural Language Understanding AI. The most challenging area in enterprise data is unstructured natural language. The process automation that we can bring to the daily business operation such as customer support and employee support would be enormous.

With this round of funding, we will accelerate our growth and take on exciting challenges:

Supervised learning as a business process

Allganize's Intent Classification and Named Entity Recognition utilize supervised learning with deep neural network and transfer learning. However, preparing high quality training data and converting user's raw input to the training data is not an easy task because tagging and maintaining the high volume of training data is labor intensive and foreign to most people.

To solve this, Allganize have introduced FAQ with personalized multi-turn conversation logic by ChatFlow and Instant Learning, which brings the almost same precision and recall of trained model with cutting the training time to 1/10th. We'll continue to push for the best practices of utilizing supervised learning, and distant supervised learning as a well-defined business process.

AI with minimal maintenance

Offline learning using the static training data, wouldn't be good enough especially for Chatbot and the voice assistant to handle dynamically changing end user's complex natural language requests day by day. However, adopting online learning using the dynamic training data as they come in would be a huge burden for the operation manager. Designing the business friendly online learning process to design AI that evolves everyday without any hassle is an important topic.

Our journey for bringing the best-in-class Natural Language Understanding AI to enterprises has just started in US, Japan and South Korea. Please stay tuned.